From the Division of Transplant Surgery (Drs Perez, McVicar, and Troppmann, and Messrs Huang and Johnson) and Transplant Medicine (Drs Gallay and Gandhi), University of California[[ndash]]Davis Medical Center, Sacramento.
Occult pretransplantation systemic inflammation will identify patients at risk for poor outcomes after renal transplantation.
Retrospective cohort study. Adhesion molecule levels were measured in pretransplantation serum samples from 86 recipients. Univariate and multivariate analyses were conducted to assess a possible correlation between serum adhesion molecule level and outcome.
University referral center.
Main Outcome Measures
Allograft rejection and survival.
Patients with low levels of vascular cell adhesion molecule 1 had less graft rejection (P = .007). Low levels of vascular cell adhesion molecule 1 independently predicted decreased rejection (relative risk, 0.17; P = .01), and high levels of vascular cell adhesion molecule 1 independently predicted graft loss (relative risk, 3.83; P = .02). Similar correlations were observed for intercellular adhesion molecule 1.
Decreased pretransplantation adhesion molecule expression correlates with less rejection, and increased levels correlate with graft loss. Assessment of pretransplantation inflammatory status may be useful in optimizing immunosuppression therapy.
OCCULT SYSTEMIC inflammation has recently been implicated as an important factor in the pathogenesis of atherosclerosis. The presence of underlying occult inflammation in apparently healthy individuals has been shown to predict the occurrence of future atherosclerotic ischemic events. Serum levels of acute-phase proteins, adhesion molecules, and cytokines have been found to be elevated in asymptomatic individuals who subsequently suffered myocardial infarction, cerebral vascular disease, and peripheral arterial disease.1- 6 Intense investigative efforts are currently underway to identify mechanisms by which inflammation is initiated, amplified, and sustained in these high-risk patients. An improved understanding of this inflammatory process will contribute to the development of strategies for preemptive targeted therapy to decrease inflammation and prevent the development of vascular disease.
There are many similarities between the molecular and cellular events involved in the pathogenesis of atherosclerosis and those that result in allograft rejection. Both processes involve leukocyte-endothelial interactions with local inflammation, endothelial activation, leukocyte adhesion, and transmigration of the vascular endothelial wall.7,8 Both processes involve cellular responses as part of innate and adaptive immunity.9,10 The occult and chronic nature of the inflammatory process involved in the pathogenesis of atherosclerosis has been relatively well studied. In contrast, few studies have explored the possibility of a similar model existing in the setting of clinical transplantation. Because of the similarities between atherosclerosis and allograft rejection, one could hypothesize that occult, systemic inflammation, potentially beginning prior to transplantation, would identify patients at high risk for allograft rejection and/or graft loss.
We have recently begun to explore the relationship between occult pretransplantation systemic inflammation and posttransplantation allograft outcome. We have demonstrated that pretransplantation systemic inflammation, manifested by elevated serum C-reactive protein (CRP) levels independently predicts which patients are at higher risk for the development of acute renal allograft rejection episodes.11 A subsequent study demonstrated that high pretransplantation CRP levels predicted eventual allograft failure.12 C-reactive protein, an acute-phase protein produced in the liver, is thought to be mainly a marker of systemic inflammation and generally not an inflammatory mediator itself. A potentially more useful approach to assessing the adverse effects of systemic inflammation might be by measurement of direct mediators of inflammation. Cellular adhesion molecules, important in the recruitment of inflammatory cells to sites of inflammation, might be ideal mediators to assess, as they are involved in inflammatory processes localized to the vascular endothelial-leukocyte interface.13 These molecules are synthesized by endothelial cells and regulate the migration of inflammatory cells out of the systemic circulation and into tissues via a multistep process involving adhesion molecules on the endothelial cells and their ligands on leukocytes. The 2 adhesion molecules most commonly studied include intercellular adhesion molecule 1 (ICAM-1) and vascular cell adhesion molecule 1 (VCAM-1). These adhesion molecules bind to leukocytes via lymphocyte function–associated antigen 1 or very late antigen 4, respectively, and can be measured systemically in soluble form. The purpose of this study was to determine if pretransplantation inflammation, as manifested by increased serum levels of soluble adhesion molecules, would identify patients at increased risk for acute rejection or graft loss after living donor renal transplantation.
Approval for this clinical study was obtained from the human subjects committee of the University of California–Davis Medical Center (Sacramento). There were 99 adults who received living related or unrelated kidney transplants at the University of California–Davis Medical Center from June 1987 to December 1998. Five of these patients had no available pretransplantation serum samples and were excluded from the study. Two patients received organs from identical twin donors and were also excluded. Six patients had incomplete posttransplantation follow-up information. Therefore, the study group includes 86 patients with available pretransplantation serum samples and clinical follow-up information.
Medical records, with the most recent follow-up information as of November 2001, were reviewed. Pretransplantation recipient variables recorded include age, sex, race, cause of end-stage renal disease, duration and type of dialysis, prior transplantation, pretransplantation blood transfusion, cytomegalovirus seropositivity status, and percentage of panel reactive antibody at the time of transplantation. Donor age, donor source (living related vs unrelated), and number of HLA mismatches were also recorded. Rejection episodes confirmed by percutaneous core needle biopsy and the cause of graft loss or death were also noted.
An aliquot from the stored serum specimen that was used for the final cytotoxicity cross-match at the time of transplantation was provided by the Sacramento Medical Foundation Blood Center (Sacramento, Calif). Specimens were frozen at −70°C until the time of assay.
Vascular cell adhesion molecule 1 and ICAM-1 assays were performed using enzyme-linked immunosorbent assay kits. All assays were done in triplicate.
Between 1987 and August 1995, the maintenance immunosuppression protocol consisted of cyclosporine, prednisone, and azathioprine. After August 1995, mycophenolate mofetil replaced azathioprine. After November 1996, 14 patients received tacrolimus, prednisone, and mycophenolate as part of a prospective randomized clinical trial. Seven patients who received identical HLA matches received cyclosporine and prednisone only.
Kaplan-Meier estimates were used to generate survival curves for the time to first episode of acute rejection (rejection-free graft survival) and overall graft survival. Those patients for whom no rejection episode was observed were considered to be censored. Separate curves were also plotted comparing time to rejection or time to graft loss based on subgroups stratified by pretransplantation levels of VCAM-1 or ICAM-1. The log-rank test was used to evaluate possible differences in rejection-free graft survival and overall graft survival between groups.
Cox regression multivariate analysis was then performed to assess whether rejection-free graft survival could be predicted by selected pretransplantation covariates. The covariates assessed included number of HLA-matched antigens, panel of reactive antigens, recipient age, history of blood transfusion, dialysis duration (pretransplantation days receiving dialysis), pretransplantation recipient cytomegalovirus serostatus, and pretransplantation levels of VCAM-1 or ICAM-1 (by quartile and as a continuous variable). All covariates were treated as continuous variables except for transfusion status (binary variable), cytomegalovirus seropositivity status (binary), and VCAM-1/ICAM-1.
Cox regression analysis was also performed to assess whether overall graft survival could be predicted by selected pretransplantation covariates. The same covariates described in the rejection-free survival multivariate analysis were also used in this analysis. One additional covariate, acute rejection, was included in this analysis to determine independent predictors of overall graft survival. The statistical analysis was performed with NCSS 2000 software (NCSS, Kaysville, Utah) and StatView 1992-1998 (SAS Institute Inc, Cary, NC). The level of statistical significance chosen was P<.05.
Table 1 presents the clinical characteristics of the study population. The mean ± SD length of follow-up, defined as the time each patient was followed up until graft loss, death, or last follow-up date, was 5.37 ± 2.55 years (range, 0.24-12.79 years).
Figure 1 shows a histogram of pretransplantation serum VCAM-1 and ICAM-1 levels. Levels of VCAM-1 ranged from a low value of 0.52 ng/mL to a high of 3.97 ng/mL (mean ± SE, 1.89 ± 0.07 ng/mL). Levels of VCAM-1 were normally distributed (Figure 1A). Levels of ICAM-1 ranged from a low of 23 ng/mL to a high of 380 ng/mL (mean ± SE, 109 ± 8.4 ng/mL). The distribution of ICAM-1 values showed a slight skewing toward lower levels (Figure 1B). A simple regression curve was constructed between the 2 adhesion molecules (Figure 2). A very significant correlation was observed between the pretransplantation serum levels of VCAM-1 and ICAM-1 (R = 0.866, P<.001).
A, Distribution of pretransplantation serum levels of vascular cell adhesion molecule 1 (VCAM-1) in 86 study patients who subsequently underwent living donor renal transplantation. B, Distribution of pretransplantation serum intercellular adhesion molecule 1 (ICAM-1) levels in 86 study patients who subsequently underwent living donor renal transplantation.
Simple regression demonstrating a significant correlation between pretransplantation serum levels of vascular cell adhesion molecule 1 (VCAM-1) and intercellular adhesion molecule 1 (ICAM-1) (R = 0.866, P<.001).
The clinical characteristics of patients with pretransplantation levels of VCAM-1 stratified by quartile are presented in Table 2. There were no significant differences in clinical characteristics among the patients in the 4 VCAM-1 quartiles.
A total of 40 (47%) of the 86 patients experienced a rejection episode during the study period, and in 48% of these patients, the episode occurred within the first 3 months. Rejection-free graft survival is defined as the time until the first rejection episode, graft loss, or death. The mean ± SE rejection-free graft survival time was 3.86 ± 3.35 years (range, 3 days to 12.79 years).
Kaplan-Meier rejection-free graft survival curves for patients, stratified by pretransplantation levels of VCAM-1, are shown in Figure 3. To determine if low pretransplantation levels of VCAM-1 identified patients who were at low risk of rejection compared with the total patient population, rejection-free graft survival was compared between VCAM-1 quartile 1 and the remainder of the patients (Figure 3A). Patients within the lowest quartile of VCAM-1 had prolonged rejection-free graft survival when compared with VCAM-1 quartiles 2 through 4 (P = .006). We then examined whether patients with the highest pretransplantation levels of VCAM-1 were at increased risk of rejection compared with the remainder of the patient population. As shown in Figure 3B, patients within the highest quartile of pretransplantation levels of VCAM-1 had no significant difference in rejection-free graft survival when compared with the lowest 3 VCAM-1 quartiles (P = .38).
Kaplan-Meier rejection-free graft survival curves for patients with pretransplantation serum levels of vascular cell adhesion molecule 1 (VCAM-1) in quartile 1 vs 2 through 4 (A) and quartiles 1 through 3 vs 4 (B).
Analysis of rejection-free graft survival, stratified according to recipient pretransplantation ICAM-1 serum levels, showed results similar to those with VCAM-1 (Kaplan-Meier curves not shown). Patients with the lowest levels of pretransplantation ICAM-1 had significantly fewer rejections when compared with the rest of the patients (ICAM-1 quartile 1 vs quartiles 2-4; P = .02). As was noted for the VCAM-1 analysis, patients with the highest pretransplantation levels of ICAM-1 had no difference in rejection-free graft survival when compared with the rest of the patient population (ICAM-1 quartiles 1-3 vs quartile 4; P = .44).
Patient survival was 95.3% (82 of 86 patients) at the end of the study period. The causes of death were sepsis (2 patients), metastatic malignancy, and trauma, occurring 3.8 to 6.6 years after transplantation. All 4 patients who died had functioning allografts at the time of death. The patient deaths were equally distributed within the VCAM-1 quartiles, with 1 death occurring in each quartile.
There were a total of 18 patients (21%) who had graft loss during the period studied, which included the 4 patients with death as the cause of graft loss. Kaplan-Meier curves of overall graft survival comparing recipients stratified by VCAM-1 quartile are shown in Figure 4. To determine if low VCAM-1 levels identified patients who were at low risk of allograft loss compared with the total patient population, overall graft survival was compared between VCAM-1 quartile 1 and the remainder of the patients (Figure 4A). Patients within the lowest quartile of VCAM-1 had a trend toward prolonged overall graft survival when compared with VCAM-1 quartiles 2 through 4 (Figure 4A) (P = .12). We then examined whether patients with the highest pretransplantation levels of VCAM-1 were at increased risk of graft loss compared with the remainder of the patient population. As shown in Figure 4B, patients within the highest quartile of pretransplantation VCAM-1 levels had decreased allograft survival when compared with the rest of the study patients (VCAM-1 quartiles 1-3 vs quartile 4; P = .02).
Overall renal allograft survival curves for patients with pretransplantation serum levels of vascular cell adhesion molecule 1 (VCAM-1) in quartile 1 vs 2 through 4 (A) and quartiles 1 through 3 vs 4 (B).
Cox regression analysis was performed to determine whether different covariates independently influenced the occurrence of allograft rejection. As presented in Table 3, of all covariates tested, only 2 independently predicted the occurrence of acute rejection. Low pretransplantation VCAM-1 levels (within the first VCAM-1 quartile) were significantly predictive of a lower incidence of allograft rejection (relative risk [RR], 0.17; P = .01). The only other covariate that independently predicted acute rejection was HLA mismatch (RR, 1.28; P = .02).
The results of the multivariate analyses to determine whether any pretransplantation variables independently predicted allograft loss are shown in Table 4. Consistent with other studies, acute rejection episodes were a very significant predictor of graft loss (RR, 8.97; P = .007). Additionally, high VCAM-1 levels independently predicted graft loss (RR for graft loss, 3.83; P = .02). When ICAM-1 replaced VCAM-1 in the multivariate analysis, the results were similar to those of VCAM-1. Specifically, a low ICAM-1 level (within the first quartile) predicted less rejection (RR, 0.22; P = .01), and an increased ICAM-1 level independently predicted graft loss (RR, 2.07; P = .06).
The findings of this study support the hypothesis that patients with increased occult pretransplantation systemic inflammation are at high risk for poor outcomes after renal transplantation. These findings, using soluble adhesion molecule levels as a measure of inflammation, are consistent with our previous work, which demonstrated that enhanced pretransplantation inflammation manifested by increased CRP levels was very predictive of renal allograft rejection and allograft loss.11,12 Other investigators have recently reported findings similar to ours. Fink et al14 have recently shown that elevated pretransplantation CRP levels may be useful in identifying patients at risk for the development of chronic allograft nephropathy. This pilot study included only 15 patients with biopsy-proven chronic allograft nephropathy; however, the adjusted risk of allograft nephropathy increased more than 3-fold with each increment in pretransplantation CRP level (by tertile). Another recent study in clinical cardiac transplantation has linked early posttransplantation systemic inflammation with an increase in arterial expression of adhesion molecules and subsequent poor long-term outcome.15 Patients with increased systemic inflammation, as manifested by elevated CRP levels during the first 3 months after transplantation, had significantly higher levels of ICAM-1 expression within the arterial endothelium, obtained by endomyocardial biopsy. These patients went on to have a more rapid development of allograft coronary artery disease and poorer allograft survival.
The cause of the inflammatory activity observed in a subset of our patients is not obvious but is consistent with a large body of recent literature that has characterized a state of chronic inflammation often seen in patients with end-stage renal disease.16- 18 The underlying causes of inflammation associated with renal failure are not completely understood. It has been postulated that the causes are multiple and may include factors related to the underlying kidney disease, the dialysis process itself, chronic infection, or to genetic factors that may regulate inflammatory cytokine expression. Clearly, a better understanding of the inflammatory state associated with end-stage renal disease is needed.
The possibility that exposure to pretransplantation hemodialysis may be one cause of inflammation that leads to poor transplantation outcomes is supported by 2 recent reports. Both investigators used the US Renal Data System to assess the effect of pretransplantation dialysis on posttransplantation outcomes. Mange et al19 very compellingly demonstrated that recipients of living donor renal allografts who had never been exposed to pretransplantation dialysis had less allograft rejection and superior allograft survival when compared with those patients who had been receiving dialysis prior to transplantation. Meier-Kriesche and Kaplan20 reported similar findings but also showed a quantitative effect in that renal allograft survival progressively worsened with increasing duration of dialysis prior to transplantation. This was true of all recipients regardless of whether they received a living donor or cadaveric allograft. A possible role of dialysis-induced inflammation as an active mechanism in these studies is presently speculative but warrants further investigation.
The main limitations to this study are due to the relatively small sample size and retrospective design. Additionally, the study period encompassed a long span of time during which changes were made in immunosuppression protocols. These changes could have influenced rates of rejection to some degree, although the effect of systemic inflammation appeared to be independent of the immunosuppressive agents used. Relative to this issue, the incidence of rejection observed with present immunosuppressive agents is substantially lower than in the recent past. It is therefore possible that the use of the present potent antirejection agents would diminish or negate the adverse effect of pretransplantation occult systemic inflammation. A prospective study involving a larger sample size and using current immunosuppressive protocols will be necessary to determine whether occult inflammation prior to transplantation leads to increased rejection and decreased allograft survival.
If these findings are confirmed in larger prospective studies, several possibilities arise that may contribute to improved outcomes in transplantation. One important focus would be the identification of patients who are at low risk for allograft rejection. The identification of these low-risk patients would make it possible to develop future protocols that use lower accumulative doses of immunosuppressive agents at the time of transplantation. These low-risk patients would probably be optimal candidates for future clinical trials using protocols aimed at the induction of immunologic tolerance. An equally important focus would be the identification of patients who have significantly increased inflammation and thus are at a high risk for rejection. A better understanding of the factors that regulate inflammation in these patients may provide a rationale for the development of pretransplantation interventions aimed at decreasing inflammation and potentially improving clinical outcomes.
In conclusion, assessment of occult pretransplantation systemic inflammation may be useful in identifying patients at high and low risk for adverse outcomes after living donor renal transplantation. Thus, assessment of pretransplantation systemic inflammatory status, using serum levels of soluble adhesion molecules, may be helpful in prospective individualization of immunosuppression therapy for renal transplantation.
Corresponding author and reprints: Richard V. Perez, MD, Department of Surgery, Division of Transplantation, University of California–Davis Medical Center, House Staff Facility, Room 2021, 2315 Stockton Blvd, Sacramento, CA 95817 (e-mail: firstname.lastname@example.org).
Accepted for publication April 18, 2002.
This study was supported in part by a University of California–Davis Health System Research Award.
This study was presented at the 74th Annual Meeting of the Pacific Coast Surgical Association; February 18, 2003; Monterey, Calif; and is published after peer review and revision. The discussion is based on the originally submitted manuscript and not the revised manuscript.
Susan L. Orloff, MD, Portland, Ore: This paper hypothesizes that occult systemic inflammation in the living donor renal transplant recipient prior to transplantation negatively impacts posttransplant outcomes as measured by acute rejection or graft loss. In this paper, the recipient inflammatory state is assessed by measuring serum levels of the soluble adhesion molecules, VCAM-1 and ICAM-1. These adhesion molecules are synthesized by endothelial cells and interact with their ligands, the α4 integrins on leukocytes, and by this interaction are important in the recruitment and migration of inflammatory cells to sites of inflammation at the vascular endothelial cell–leukocyte interface. I have 4 points of discussion for the authors:
Can you discuss the proposed mechanism of this pretransplant inflammation in the recipient and the posttransplant rejection and graft loss? Is this a nonspecific phenomenon or is it related to specific factors, such as chronic occult infection, chronic dialysis, the native renal disease, and/or genetic factors? You have measured soluble adhesion molecule levels in the recipients in this study. Have you similarly measured levels in the living donors of these allografts? This would be important given that the endothelial cells of the graft are donor derived and it is at this cellular level that the receptor-ligand interactions of the adhesion molecules occur, resulting in tissue inflammation and damage.
You remark on the similarities between the molecular and cellular events involved in the pathogenesis of atherosclerosis and those resulting in allograft rejection. Are these similarities not more between atherosclerosis and chronic vascular rejection, and not acute rejection, which is one of your end points of study? Can you comment on this?
The rejection rates reported in your study of 47% during the entire study, 48% of which were in the first 3 months, are significantly higher than the current national average of about 10% to 15% rate of rejection in the first year. Given these findings, is the study relevant in today's setting of potent and very effective immunosuppressive regimens?
Given your findings, are you measuring pretransplant serum soluble adhesion molecules in your patients, and accordingly, altering your posttransplant immunosuppressive regimens? In other words, have you put your data to test prospectively? What interventions have you thought of to decrease inflammation prior to transplantation in order to improve posttransplant outcomes? Along this vein, in the January 2003 issue of The New England Journal of Medicine, there are 2 reports on very favorable results of clinical trials using natalizumab, a recombinant monoclonal antibody against α4 integrins, for the treatment of Crohn's disease and multiple sclerosis.
Nataluzimab has therapeutic effects because it blocks the ability of the integrin α4b1 to bind to its endothelial counter-receptor VCAM-1 and prevents lymphocyte entry into tissues and consequent inflammation and tissue damage. Have you thought of some such similar treatment for transplant recipients?
Chris E. Freise, MD, San Francisco, Calif: This is a great contribution to the growing field of immune monitoring that all of us as transplant surgeons are hopeful will continue to improve, so that we can better tailor drugs towards our patients. I have one question related to an experimental observation that we have made. We are actually interested in measuring ICAM-1 tissue levels in stored rat kidneys and have found a correlation between cold storage time and increased ICAM-1 tissue levels in stored kidneys and have hypothesized that there may be a relationship between ICAM-1 tissue levels and subsequent delayed graft function. Realizing that this is a live-donor population, you probably had a relatively low rate of delayed graft function. However, I was curious if you saw any correlation between delayed graft function or slow graft function and your inflammatory markers.
Dr Perez: In response to the very insightful questions by Dr Orloff, first, regarding the mechanisms of inflammation, it has been well documented that patients with end-stage renal disease are patients who have high levels of inflammation. They have high levels of acute-phase reactants, such as C-reactive protein, adhesion molecules, fibrinogen, and cytokines, namely IL-6. The mechanism for this is multifactorial. There is much literature suggesting that the dialysis process itself is inflammatory, that the native kidney disease may have an effect, and chronic infections, both viral and bacterial, very analogous to what we see in the cardiac literature where there is evidence that chronic CMV, Chylamidia infection, and herpesvirus may contribute to inflammation. So, certainly, we need to understand more regarding the cause of the inflammation in these patients in order to intervene.
The idea about looking at inflammation in the donor is very relevant. We haven't done that in our patients but that certainly bears investigating. It has been shown in cadaveric transplantation that definitely, inflammatory events due to brain death in cadaveric donors have a very significant and profound effect on graft outcome. The benefit we see in living donation may be so striking because the donors aren't subjected to the same changes in inflammation associated with brain death. But that is certainly something that we should look at.
The high rates of rejection observed here are reflective of our old immunosuppressive protocols. A rejection rate of 30% to 40% a decade ago was pretty much normal. And whether this is relevant today, it is very relevant because even though the rejection rates are lower today, it's due to the very potent immunosuppressive agents that we have, and there is still a role for identifying patients that are at high risk so we can tailor the immunosuppression appropriately.
What are we doing now? First we are doing prospective confirmation of these findings so in today's current immunosuppressive therapy, we are looking at a whole panel of inflammatory mediators, including C-reactive protein, certain cytokines, and adhesion molecules. We would like to take this in 2 different directions. If we can identify patients who are at low risk for rejection by such a panel, then we could treat those patients with very minimal immunosuppression—maybe no steroids or a number of interventions. Alternatively, if we identify patients who are at very high risk for rejection, then we might have the opportunity to intervene as has been done in the atherosclerotic literature using some of the similar agents, like the statins, or agents directed specifically at adhesion molecules.
Dr Freise brought up a very good point in terms of looking at the correlation for inflammation and delayed graft function. We didn't do that in this patient population, but we do have some preliminary evidence in our cadaveric patients, recipients of cadaveric organs, that inflammation as measured by elevated C-reactive protein actually may predict higher risk for delayed graft function. So there is certainly an inflammatory process ongoing in the recipient that that may lead to increased expression of adhesion molecules or other inflammatory mediators which will cause graft dysfunction, and we are seeing that in our cadaveric patients.
Thank you for submitting a comment on this article. It will be reviewed by JAMA Surgery editors. You will be notified when your comment has been published. Comments should not exceed 500 words of text and 10 references.
Do not submit personal medical questions or information that could identify a specific patient, questions about a particular case, or general inquiries to an author. Only content that has not been published, posted, or submitted elsewhere should be submitted. By submitting this Comment, you and any coauthors transfer copyright to the journal if your Comment is posted.
* = Required Field
Disclosure of Any Conflicts of Interest*
Indicate all relevant conflicts of interest of each author below, including all relevant financial interests, activities, and relationships within the past 3 years including, but not limited to, employment, affiliation, grants or funding, consultancies, honoraria or payment, speakers’ bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued. If all authors have none, check "No potential conflicts or relevant financial interests" in the box below. Please also indicate any funding received in support of this work. The information will be posted with your response.
Register and get free email Table of Contents alerts, saved searches, PowerPoint downloads, CME quizzes, and more
Subscribe for full-text access to content from 1998 forward and a host of useful features
Activate your current subscription (AMA members and current subscribers)
Purchase Online Access to this article for 24 hours
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited: 6
Customize your page view by dragging & repositioning the boxes below.
and access these and other features:
Enter your username and email address. We'll send you a link to reset your password.
Enter your username and email address. We'll send instructions on how to reset your password to the email address we have on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.