HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:
Guest Access | Sign In via User Name/Password

This Article Abstract Full Text (PDF) Free Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Article Archive Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (12) Citing Articles via Google Scholar Google Scholar Articles by Clermont, G. Articles by Pinsky, M. R. Search for Related Content PubMed PubMed Citation Articles by Clermont, G. Articles by Pinsky, M. R.

Does Acute Organ Dysfunction Predict Patient-Centered Outcomes?^*

^* From the Clinical Research, Investigation, and Systems Modeling of Acute Illness Laboratory (Drs. Clermont and Angus), Department of Critical Care Medicine (Dr. Pinsky), and Department of Medicine (Dr. Fine), Division of General Internal Medicine, University of Pittsburgh, Pittsburgh, PA; and Health Process Management, LLC (Mr. Linde-Zwirble and Mr. Griffin), Doylestown, PA.

Correspondence to: Derek C. Angus, MD, MPH, 604 Scaife Hall, University of Pittsburgh, Department of Anesthesiology/Critical Care Medicine, 200 Lothrop St, Pittsburgh, PA 15213; e-mail: angusdc{at}anes.upmc.edu

	Abstract

TOP Abstract Introduction Materials and Methods Results Discussion References

 
Background: Long-term patient-centered outcomes after acute illness may be associated with baseline health status, the development of acute organ dysfunction (AOD), or both.

Study objective: To determine whether AOD (occurring in the first 30 days) was independently associated with 90-day survival, functional status, and health-related quality of life (HRQL) after controlling for baseline health status in patients who were hospitalized with community-acquired pneumonia (CAP) and survived to day 30.

Design: Prospective observational study.

Setting: Four hospitals in Pennsylvania, Massachusetts, and Nova Scotia, Canada, between October 1991 and March 1994.

Patients: One thousand three hundred thirty-nine patients who were hospitalized with CAP.

Interventions: Baseline and 90-day quality-of-life and functional status questionnaires.

Measurements and results: We determined the 90-day survival rate in all patients (n = 1,339) and the functional status and HRQL in subsets of 261 and 219 patients, respectively. AOD occurred in one or more organ system in 639 patients (47.7%) and in two or more organ systems in 255 patients (19.1%). In univariate analyses, greater AOD was associated with a higher mortality rate (p < 0.0001), a lower HRQL (p = 0.006), and lower functional status (p = 0.009) at 90 days. However, after adjusting for baseline HRQL, AOD was not associated with mortality (p = 0.47) or HRQL (p = 0.14) at 90 days and was only weakly associated with 90-day functional status (p = 0.02).

Conclusions: Although patients who develop AOD are at risk for late adverse outcomes, their risk is due predominantly to poor baseline status prior to illness and not to the organ dysfunction per se. Therefore, AOD does not appear to have significant long-term ramifications for patient-centered outcomes.

Key Words: community-acquired pneumonia • critical illness • mortality • organ dysfunction • organ failure • outcome • quality of life • surrogate outcomes

	Introduction

TOP Abstract Introduction Materials and Methods Results Discussion References

 
Based on the observation that acute organ dysfunction (AOD) is associated with increased short-term mortality rates,^{1 2 3 4} it is possible that short-term survivors who develop AOD will have worse subsequent medical outcomes than those who did not develop AOD. There is also evidence suggesting that baseline health status prior to acute illness is the most important predictor of long-term outcome in those who survive the critical phase.^{5 6 7} A better understanding of the relative importance of these two factors, AOD and baseline health status, for subsequent outcome may aid in patient decision making. However, ascertaining prior health status, especially health-related quality of life (HRQL), is often difficult in studies of acutely ill patients.

We assessed the relationship between the development of AOD in the first 30 days following the onset of an acute illness and subsequent outcome in a large cohort of patients hospitalized with community- acquired pneumonia (CAP). Specifically, we tested the hypothesis that a greater degree of AOD in those patients who were alive at 30 days was associated with an increased mortality rate at 90 days, a delayed return to usual activities, and a deterioration in functional status and quality of life. We selected patients with CAP because they represent a large, relatively homogenous population who, although at high risk of developing AOD, are often well enough when first seen to allow an assessment of their baseline HRQL and functional status prior to the onset of acute illness.

	Materials and Methods

TOP Abstract Introduction Materials and Methods Results Discussion References

 
Study Population
We studied all hospitalized patients in the Pneumonia Patient Outcomes Research Team prospective cohort⁸ for whom chart review was complete and 90-day outcome was recorded (n = 1,339). Patients were enrolled at four hospitals in Pennsylvania, Massachusetts, and Nova Scotia, Canada, between October 1991 and March 1994. Potential study subjects were identified by research assistants through daily reviews of hospital admissions and radiology department records and logs of patients presenting to emergency departments.

Inclusion in the study required that patients be 18 years of age, have one or more acute clinical symptoms suggestive of pneumonia, have radiographic evidence of pneumonia within 24 h of presentation (as determined by a local study site radiologist), and provide informed consent (by proxy where necessary) for baseline and follow-up interviews. Patients ineligible for the study included those who had been discharged from an acute care hospital within the prior 10 days, those who had been enrolled previously, those who had received clinical diagnoses of AIDS or who were known to be seropositive for HIV, and those with a definitive diagnosis other than pneumonia as the likely explanation for their pulmonary symptoms and the presence of infiltrates. The baseline assessment included sociodemographic characteristics, physical examination, laboratory data, employment status, and determination of the presence of pneumonia-specific symptoms and chronic diseases. Further details on this cohort have been provided elsewhere.⁹

Classification of AOD
We ascertained the incidence of AOD by detailed chart review until hospital discharge or a maximum of 30 days of hospitalization. AOD was assessed on a scale of 0 to 4 for the following six organ systems: cardiovascular; respiratory; neurologic; renal; hematologic; and GI. This assessment was based on defined a priori criteria (Table 1 ), in which a score was assigned if any of the criteria were met at any point during the 30-day chart review period. The rationale for the choice of organ systems was based on published organ dysfunction scores^{3 4} and had been determined prior to the current study. We dichotomized AOD as present or absent, as described in Table 1 (ie, "Base Definition" column). We stratified the study population into three subgroups on the basis of acute dysfunction occurring in zero, one, or more than one organ system.

Patient-Centered Outcomes
Mortality Rate We measured short and intermediate-term mortality. As a measure of short-term outcome, we measured the 30-day mortality rate across cohorts of patients with acute dysfunction in zero, one, or more than one organ system. As a measure of intermediate-term outcome, we observed all patients who survived 30 days until death, or a maximum of 90 days, and calculated the 90-day mortality rate.

HRQL
We assessed HRQL using the short form Medical Outcomes Survey 36-item short form (SF-36) health survey¹⁰ at baseline and 90 days in a convenience cohort of 377 inpatients. Initially, we restricted this assessment to patients classified as being at low risk of death because we believed that patients at greater risk of death would be too sick to complete the quality-of-life assessment (n = 274).¹¹ As the study progressed, patients who were more sick were included as well (n = 103). We administered all questionnaires by direct interview with the patient or a proxy, by telephone or by written completion of a mailed version of the questionnaire. We included in our final analysis only patients who completed both the baseline and 90-day HRQL questionnaires (n = 219). We assessed baseline HRQL on admission to the hospital using the 4-week recall version of the SF-36 and used the 1-week recall survey to assess HRQL at 90 days.¹²

The SF-36 score is segregated into physical and mental component summary measures. Because acute illness can influence perceived physical and mental states differently, we analyzed physical and mental component summary scores separately. We used the summary scores, which had been standardized previously to a mean of 50 (SD, 10), in a general healthy population.¹²

Functional Status
We assessed functional status in the same subset of patients by inquiring about 15 activities performed on a daily basis (ie, walking, getting out of bed, feeding, bathing, dressing, grooming, going to the bathroom, sphincter control, ability to make telephone calls, preparing meals, housekeeping, shopping, self-administering medications, handling finances, and driving or riding public transport). These activities are an expanded version of the activities-of-daily-living index published by Katz et al.¹³ Activities were graded according to a 4-point system in which performance of the activity "without help" scored 1, performance "with a little help" scored 2, performance "with a lot of help" scored 3, and the response "do not do activity," scored 4 points. Some activities did not include the last choice. We generated a summary score (range, 15 to 52 [a score of 15 signifies autonomous function in all recorded activities]) as the sum of the scores across all 15 activities. Two hundred sixty-one patients completed both the baseline and 90-day functional status questionnaires.

Return to Work or Usual Activities
We included work status as part of the baseline demographic data on all patients. Baseline level of activity, time of return to the usual level of activity, as well as return to work for people employed prior to the onset of pneumonia, were part of a questionnaire that we administered to patients (or their proxies), 801 of whom (60%) were alive at 90 days and responded to a follow-up questionnaire.

Statistical Analysis
We compared rates across cohorts using the ² test. We compared Kaplan-Meier survival curves for subgroups with different degrees of AOD and for proportions of patients resuming usual activities using the log-rank test.^{14 15}

We compared SF-36 summary scores and functional status scores at each time point by one-way analysis of variance (ANOVA). Post hoc comparisons of scores across organ failure subgroups were performed using the Bonferroni procedure. We also compared physical and mental component summary scores within subgroups pairwise among baseline, 30-day, and 90-day scores by two-tailed paired t tests.

The incidence of organ dysfunction was sufficient to provide 80% power to detect a 6-point difference in SF-36 component summary scores between the general reference population and patients with dysfunction in zero, one, or more than one organ system. There was also 80% power to detect a 5-point difference between the baseline scores and the 90-day scores in the smallest subgroup (ie, organ failure in more than one organ system) and 80% power to detect an 8-point difference between subgroups with different degrees of organ dysfunction at either baseline or 90 days.

We constructed regression models to determine whether AOD predicted 90-day mortality rates in 30-day survivors controlling for age. We also constructed regression models, controlling for baseline scores and age, to determine whether AOD predicted the 90-day SF-36 physical and mental components scores, and 90-day functional status summary score. These models were constructed using the three different methods of assessing the severity of organ dysfunction described above. Database manipulation was conducted using appropriate software (Visual FoxPro, version 6.0; Microsoft Corp; Redmond, WA), and statistical analyses were likewise conducted with an appropriate software package (SPSS, version 10.1; SPSS, Inc; Chicago, IL). For ANOVA, t tests, and ² tests, a p value < 0.05 was considered to be significant. Post hoc comparisons among the three organ failure subgroups were performed by a Bonferroni procedure, assuming significance at p < 0.0167. The mean summary scores are presented with their 95% confidence intervals (CIs).

	Results

TOP Abstract Introduction Materials and Methods Results Discussion References

 
Baseline demographic characteristics are presented in Table 2 . The average hospital length of stay for the entire cohort was 10.9 days (median, 8 days; interquartile range, 5 to 12 days). Only 4.5% of patients had lengths of stay exceeding 30 days. The incidence of AOD in the first 30 days was 47.7% (n = 639), with 19.0% (n = 255) of the cohort developing dysfunction in more than one organ system. Of the patients who manifested organ dysfunction during the first 30 days of their hospital stay, 21.4% (n = 286) had acute respiratory dysfunction, 16.5% (n = 221) had renal dysfunction, 14.0% (n = 188) had GI or hepatic dysfunction, 13.0% (n = 174) had hematologic dysfunction, 12.1% (n = 162) had cardiovascular dysfunction, and 2.7% (n = 36) had neurologic dysfunction. The respiratory and hematologic systems were most often the sole dysfunctional organ systems. Not surprisingly, there were strong associations between the development of AOD and the 30-day mortality rate, ICU and hospital length of stay, and ICU admission rate. The 30-day mortality rates were 3.3%, 7.8%, 13.3%, 25.8%, and 39.1%, respectively, in patients with acute dysfunction in zero, one, two, three, or four or more organ systems (p < 0.0001). Hospital mean length of stay varied from 8.4 to 30.7 days (p < 0.0001), and ICU admission rate varied from 2.1 to 71.7% (p < 0.0001) across the same subgroups.

View larger version (11K): [in this window] [in a new window] [Download PPT slide]   Figure 1.. Kaplan-Meier survival analysis by degree of AOD. Among patients surviving at least 30 days, the cohort with acute dysfunction in more than one organ system had significantly worse survival times (p < 0.05 [log rank test]).

View larger version (25K): [in this window] [in a new window] [Download PPT slide]   Figure 2.. SF-36 summary scores before the onset of CAP, at 30 days, and at 90 days reported by degree of AOD. Top, A: Physical component summary scores. Bottom, B: mental component summary scores. Physical component summary scores were different from baseline at 30 days (*p < 0.05 [paired t test]) and did not completely recover by 90 days. Mental component summary scores recovered to baseline levels by 30 and 90 days. There was a trend toward a difference across subgroups with different degrees of organ dysfunction at 90 days (p = 0.04 [ANOVA]). Organ dysfunction = AOD.

Functional Status
Three hundred thirty patients completed the baseline functional status questionnaire. Of the 261 survivors who also completed the questionnaire at 90 days, 57.5%, 30.2%, and 15.4%, respectively, had AOD in zero, one, or more than one organ system. Baseline functional status, as measured by the ability to perform 10 of the 15 specific activities (ie, getting out of bed, walking, bathing, dressing, grooming, prepare meals, housekeeping, shopping groceries, driving, and using public transportation) decreased with the degree of AOD. Summary scores increased (reflecting poorer functional status) in patients with AOD in zero, one, and more than one organ system (19.2, 21.4, and 23.9, respectively; p < 0.002). At 90 days, the scores for seven specific activities (ie, sphincter control, preparing meals, housekeeping, shopping groceries, handling finances, driving, and using public transportation) and summary scores were significantly higher than baseline scores in subgroups with AOD in one organ system (22.5 vs 21.8 for summary score; p < 0.05) or more than one organ system (26.2 vs 23.8 for summary score; p < 0.05). In multivariate regression analysis, baseline functional status was strongly associated with 90-day functional status (p < 0.001), and AOD was weakly associated with 90-day functional status (p = 0.02).

Return to Usual Activities or Work
Of the 1,339 study patients, follow-up information on their return to usual activities was available in 882 (59.3%). The median times to return to usual activities were 18 days (95% CI, 16 to 20), 23 days (95% CI, 20 to 26), and 59 days (95% CI, 18 to 100), respectively, after presentation for patients with AOD in zero, one, and more than one organ system (p = 0.002 between those patients with AOD in zero or one organ system; p < 0.0001 between those patients with AOD in one and more than one organ system) [Fig 3 ]. Information on the return to usual activities was unavailable for about one third of patients known to be alive at 90 days. Of the 33 patients known to have returned to usual activities after 90 days, 10, 12, and 11, respectively, had AOD in zero, one, and more than organ system. Because there was a high censoring rate and the censoring was potentially biased, we could not estimate with confidence the 90-day rate of return to usual activities.

View larger version (14K): [in this window] [in a new window] [Download PPT slide]   Figure 3.. Kaplan-Meier analysis of the return to usual activities by degree of organ dysfunction. Information on the return to usual activities was available for two thirds of the patients known to be alive at 90 days. Patients with a greater degree of organ dysfunction resumed usual activities later (p < 0.001 [log rank test]). Because of the high rate of censoring and the potential in bias as to the rate of return to activities in those censored, the absolute rates of return at 90 days cannot be estimated with confidence.

	Discussion

TOP Abstract Introduction Materials and Methods Results Discussion References

 
We demonstrated that patients developing AOD had poorer long-term outcomes and slower recovery times. However, after controlling for baseline health status, AOD was not independently predictive either of subsequent mortality or of HRQL and was only weakly associated with subsequent functional status. These data suggest that patients who survive acute organ failure are as likely to regain their prior health state as those who survived the same acute illness without developing organ failure.

At first, our findings may appear to be counterintuitive. There is good evidence that increasing degrees of AOD are associated with a greater risk of in-hospital or short-term mortality.^{2 3 4 16 17} Our study confirms this association. However, the focus of our study was not on those who succumb in the first month but on those who survive. Specifically, we have tried to reconcile findings of previous studies that AOD portends a worse short-term prognosis with findings from other studies that baseline functional status is the predominant risk factor for long-term patient-centered outcomes.^{5 6 7} The new finding in our study is that patients who develop organ failure are more likely to have had worse health status prior to acute illness, and it is this prior status that is the strongest predictor of long-term outcome in those who survive initially.

Perhaps our findings serve only to quantify a well-known principle of good clinical care: that a thorough understanding of a patient’s medical and social history is key to good care and decision making. We believe our data can help clinicians to inform patients and their families that, should the patient survive the acute illness, he or she has an excellent chance of returning to his or her baseline health status. Such information may well be helpful in decisions regarding the commitment to life- supporting therapies for organ support.

Given the importance of baseline functional status and HRQL as predictors of subsequent outcome after acute illness, we encourage investigators to measure these factors in studies of the acutely ill, especially in nonrandomized trials in which the uneven distribution of these variables could be an important confounder. This study also has implications for the use of organ failure scores in clinical trials. Organ failure scores may play a very useful role in describing the magnitude of a patient’s acute illness and indicating the need for hospital resources and the risk of in-hospital death. However, because acute organ failure was not independently associated with differences in long-term outcomes, the significance of a therapy targeted at an acute illness (eg, an antisepsis therapy) that reduced organ failure without changing short-term mortality rates should be interpreted with caution.

There are limitations to this study. First, we focused on one condition (CAP) and do not know whether our findings will extend to other acute illnesses. However, CAP is the third most common condition requiring hospitalization in adults,¹⁸ and a substantial proportion of patients with CAP require ICU care. Second, we only conducted HRQL assessments in a nonrandom subset of our entire cohort, which had less organ failure than the overall cohort. Within this subset, we retained adequate power to detect meaningful differences in HRQL between patients with different amounts of organ failure. For example, we were powered to find an 8-point difference in the SF-36 score, which is comparable to the difference between well-controlled and poorly controlled diabetes mellitus.¹⁹ However, it is possible that smaller, yet functionally significant, differences existed between patients with different degrees of organ dysfunction. Third, we used an arbitrary cutoff point for the definition of the presence of organ dysfunction. To investigate the potential impact of choosing different cut-points, we repeated the analysis using different cutoff points with identical conclusions. We also could not quantify organ dysfunction on the basis of the number of days with organ dysfunction. Such a measure might have resulted in a better quantification of the burden of organ dysfunction. Fourth, we quantified organ dysfunction using a scheme designed primarily for ICU patients, yet most patients were not admitted to the ICU. Less frequent physiologic assessment in hospital floor patients may result in lower organ dysfunction scores. Such a bias might have caused a smaller difference in scores between subgroups, however, this was unlikely to modify conclusions, according to the simulations we conducted (data not shown). Fifth, we only followed patients to 90 days. While this is longer than many studies of the acutely ill, significant changes in patient-centered outcomes may occur beyond this time. Nevertheless, our study is one of the first and largest studies to document HRQL both before and after acute illness and, as such, provides an important reference point for future studies of HRQL in other acute illnesses, perhaps over longer time horizons.

	Footnotes

 
Abbreviations: ANOVA = analysis of variance; AOD = acute organ dysfunction; CAP = community-acquired pneumonia; CI = confidence interval; HRQL = health-related quality of life; SF-36 = Medical Outcomes Survey 36-item short form

This research was funded by the Agency for Health Care Policy and Research (grant R01 HSO 6468) as part of the Pneumonia Patient Outcomes Research Team, by the National Institute of General Medical Sciences (Grant No. R01 GM 061992-01), and by Amgen, Inc (Thousand Oaks, CA).

Received for publication December 14, 2000. Accepted for publication December 12, 2001.

	References

TOP Abstract Introduction Materials and Methods Results Discussion References

 

1. Hebert, PC, Drummond, AJ, Singer, J, et al (1993) A simple multiple system organ failure scoring system predicts mortality of patients who have sepsis syndrome. Chest 104,230-235[Abstract/Free Full Text]
2. Le Gall, JR, Klar, J, Lemeshow, S, et al (1996) The logistic organ dysfunction system: a new way to assess organ dysfunction in the intensive care unit; ICU Scoring Group. JAMA 276,802-810[Abstract]
3. Marshall, JC, Cook, DJ, Christou, NV, et al (1995) Multiple organ dysfunction score: a reliable descriptor of a complex clinical outcome. Crit Care Med 23,1638-1652[CrossRef][ISI][Medline]
4. Knaus, WA, Draper, EA, Wagner, DP, et al (1985) Prognosis in acute organ-system failure. Ann Surg 202,685-693[ISI][Medline]
5. Vazquez Mata, G, Rivera Fernandez, R, Gonzalez Carmona, A, et al (1992) Factors related to quality of life 12 months after discharge from an intensive care unit. Crit Care Med 20,1257-1262[ISI][Medline]
6. Broslawski, GE, Elkins, M, Algus, M (1995) Functional abilities of elderly survivors of intensive care. J Am Osteopath Assoc 95,712-717[Abstract]
7. Wu, AW, Damiano, AM, Lynn, J, et al (1995) Predicting future functional status for seriously ill hospitalized adults: the SUPPORT prognostic model. Ann Intern Med 122,342-350[Abstract/Free Full Text]
8. Fine, MJ, Auble, TE, Yealy, DM, et al (1997) A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 336,243-250[Abstract/Free Full Text]
9. Fine, MJ, Stone, RA, Singer, DE, et al (1999) Processes and outcomes of care for patients with community-acquired pneumonia: results from the Pneumonia Patient Outcomes Research Team (PORT) cohort study. Arch Intern Med 159,970-980[Abstract/Free Full Text]
10. Ware, JE, Jr, Sherbourne, CD (1992) The MOS 36-item short-form health survey (SF-36); I: Conceptual framework and item selection. Med Care 30,473-483[ISI][Medline]
11. Fine, MJ, Hanusa, BH, Lave, JR, et al (1995) Comparison of a disease-specific and a generic severity of illness measure for patients with community-acquired pneumonia. J Gen Intern Med 10,359-368[ISI][Medline]
12. Ware, JE, Jr, Kozinski, M, Keller, SD (1994) SF-36 physical and mental health summary scales: a user’s manual. The Health Institute Boston, MA.
13. Katz, S, Downs, TD, Cash, HR, et al (1970) Progress in development of the index of ADL. Gerontologist 1,20-30
14. Kaplan, EL, Meier, P (1958) Nonparametric estimation from incomplete observations. J Am Stat Assoc 53,457-481[CrossRef][ISI]
15. Norman, GR, Streiner, DL (1993) Biostatistics: the bare essentials. ,190-191 Mosby Year Book St. Louis, MO.
16. Vincent, JL, Moreno, R, Takala, J, et al (1996) The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure: on behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med 22,707-710[ISI][Medline]
17. Zimmerman, JE, Knaus, WA, Wagner, DP, et al (1996) A comparison of risks and outcomes for patients with organ system failure: 1982–1990. Crit Care Med 24,1633-1641[CrossRef][ISI][Medline]
18. Popovic, JR, Hall, MJ (2001) 1999 National hospital discharge survey: advance data report No. 319. National Center for Health Statistics Hyattsville, MD.
19. Lydick, EG, Epstein, RS (1996) Clinical significance of quality of life data. Spilker, B eds. Quality of life and pharmaeconomics in clinical trials ,461-465 Lippincott-Raven Philadelphia, PA.

This article has been cited by other articles:

				T. Dremsizov, G. Clermont, J. A. Kellum, K. G. Kalassian, M. J. Fine, and D. C. Angus Severe sepsis in community-acquired pneumonia: when does it happen, and do systemic inflammatory response syndrome criteria help predict course? Chest, April 1, 2006; 129(4): 968 - 978. [Abstract] [Full Text] [PDF]

				D. C. Angus, T. J. Marrie, D. S. Obrosky, G. Clermont, T. T. Dremsizov, C. Coley, M. J. Fine, D. E. Singer, and W. N. Kapoor Severe Community-acquired Pneumonia: Use of Intensive Care Services and Evaluation of American and British Thoracic Society Diagnostic Criteria Am. J. Respir. Crit. Care Med., September 1, 2002; 166(5): 717 - 723. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Free Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Article Archive Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (12) Citing Articles via Google Scholar Google Scholar Articles by Clermont, G. Articles by Pinsky, M. R. Search for Related Content PubMed PubMed Citation Articles by Clermont, G. Articles by Pinsky, M. R.