|
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |
|
|
|
|||||||||
|
|||||||||||
|
J Am Coll Cardiol, 2004; 43:624-628, doi:10.1016/j.jacc.2003.11.017 © 2004 by the American College of Cardiology Foundation |
,*
Cardiovascular Imaging Research Core Laboratory, Vancouver Hospital, University of British Columbia, Vancouver, Canada
* Reprint requests and correspondence: Dr. G. B. John Mancini, Vancouver Hospital, 3300-950 West 10th Avenue, Vancouver, British Columbia, Canada V5Z 4E3.
mancini{at}interchange.ubc.ca
Previous investigations are summarized in Table 1, which is organized according to the method used to measure endothelial dysfunction. The investigations undertaken at the time of coronary angiography and those using venous occlusion plethysmography in conjunction with intra-brachial injection of test agents represent important, "proof of concept" studies (2–8). Those using brachial artery cuff occlusion and ultrasound to measure flow-mediated dilation (9–14) investigate an approach with a potentially broad clinical applicability and use the methodology most comparable with the study from the University of Queensland (1). All of these works, including the current one, conclude that measures of endothelial vascular dysfunction are of prognostic importance. The main issue is whether this information is additive to currently available testing methods, particularly imaging tests.
|
Perticone et al. (7) and Heitzer et al. (8) used the method of forearm venous plethysmography. Although both groups concluded that measures of endothelial dysfunction were of prognostic importance, only Heitzer et al. (8) indicated that this was independent of the number of significantly diseased vessels based on angiography.
There are significant limitations in similar studies that used completely non-invasive methods (9–13). For example, Gokce et al. (10,11) and Brevetti et al. (12) studied only patients with peripheral vascular disease. Modena et al. (13) studied only post-menopausal, hypertensive females with endothelial dysfunction at baseline. None incorporated other imaging-based, risk stratification methods. The study by Neunteufl et al. (9), based on only 73 patients, actually found no independent prognostic importance of flow-mediated dilation measurement when angiographic coronary vessel disease (>30% diameter stenosis) was incorporated into a multivariate analysis. Chan et al. (14) evaluated the prognostic importance of flow-mediated dilation in the context of extensive metabolic tests, exercise stress tests, carotid ultrasound measures of carotid plaque, and intima medial thickness. The non-invasive measurement of endothelial dysfunction was the dominant, independent predictor of outcome in a cohort of coronary patients enrolled in an intensive, cardiac rehabilitation program. Carotid ultrasound measures of atheroma burden also were predictive. Even more important, endothelial dysfunction appeared to modulate the prognostic impact of the degree of carotid atheroma burden. That is, the worst prognosis was seen in those patients with severe grades of endothelial dysfunction concomitant with a high degree of carotid plaque. Accordingly, both vascular structure and function appeared to be important. However, the study consisted of only 152 patients.
The current study (1) adds in many important ways to this relatively small body of literature. Excluding a predominantly retrospective analysis (6), the current work represents the largest cohort of patients (444 subjects) of any of the studies. It used non-invasive measures of endothelial dysfunction; therefore, the results are potentially applicable in a broad way. The follow-up time period and event rates were substantial enough to allow assessment of predictive power with respect to total mortality, cardiovascular death, and cardiovascular events. It incorporated the Framingham models, thereby providing a way to link the results to a common method used to stratify global risk. Most importantly, the authors incorporated sophisticated, multivariate analyses to tease out independent predictors of risk from an extensive battery of commonly available tests, in addition to endothelial dysfunction measurement. In this regard, the work is similar in methodology to that of Chan et al. (14), and they represent the only studies that incorporate carotid ultrasound measurements.
Despite these strengths, the conclusion that endothelial dysfunction measurement is not independently predictive of outcome is at variance with many of the preceding works. It is also at variance with general experience in other cardiovascular arenas. For example, for any given ejection fraction, patients with a higher functional capacity have a better prognosis. For any given degree of coronary disease, those with inducible ischemia fare worse. Is the current literature, which is dominated by articles suggesting that endothelial dysfunction is an independent predictor of prognosis, merely a reflection of publication bias, or are there reasons to mollify the negative conclusions? The patient cohort is definitely different from those of previous studies. Based on Table 1 of the article (1), fully 75% of the cohort had renal dysfunction, of which a substantial proportion were on dialysis and/or transplanted. The risk profile and the battery of tests that might be useful in this cohort may be substantially different than in a more typical population. This large subgroup is very likely to have had an advanced degree of atherosclerotic burden and left ventricular hypertrophy that would not be expected in a more general population. Even the subgroup of "multiple risk factor patients" included 132 patients (over 80% of the subgroup) with either chronic renal impairment or previous kidney transplantation. Almost 30% of the entire cohort and 21% of the "multiple risk factor" subgroup were diabetics, representing yet another group with a high likelihood of more advanced structural disease of the vasculature. Furthermore, 14% of the cohort had very advanced coronary disease not amenable to revascularization. These concerns about the nature of the patient population are strengthened by the finding that the actual event rates in the cohort overall and in all the subgroups were greater than predicted by the Framingham algorithms. Indeed, in the dialysis patients without previous myocardial infarction and in the "multiple cardiovascular risk factor" group, the actual event rates were higher than predicted by factors of 5.6 and 2.2, respectively. Accordingly, measures of vascular endothelial function, often present long before structural disease, may not be useful in cohorts already at very elevated risk and in whom structural abnormalities such as intima medial thickness and left ventricular hypertrophy are already well entrenched.
A second reason for mollifying the conclusions of the study pertains to the absence of functional testing, other than flow-mediated dilation testing, in a large component of the cohort. Nearly 40% of the cohort did not undergo stress testing. In the subanalysis that was based on 277 patients who did undergo stress testing, it is fascinating to note that the resulting, independent, image-based predictors of events were the measurements of endothelial dysfunction, intima medial thickness, and left ventricular hypertrophy. These results were not emphasized by the authors, but they are remarkably concordant with those of Chan et al. (14).
The work also shares several limitations with most of the previous studies in this field. It is well known that endothelial dysfunction can be treated with several classes of agents, notably angiotensin-converting enzyme inhibitors and statins among many other classes. The authors have no information on how the prognostic importance of initial and/or serial measures of endothelial function might have been altered by changes in treatment occurring during the long follow-up period. Modena et al. (13) demonstrated that a beneficial response to antihypertensive therapy and that a subsequent, low event rate could be predicted based on whether endothelial dysfunction improved or not within the first six months of treatment. Similarly, Chan et al. (14) demonstrated that outcome was predicted not solely by the degree of baseline endothelial dysfunction measurement but also by whether intensive, secondary prevention treatment resulted in an improvement or not of endothelial dysfunction. In contrast, carotid ultrasound changes were not predictive within this timeframe. If these findings were substantiated, it would suggest that treatment intensification or modification could be guided by measuring endothelial function serially. This gauge of therapy could be obtained long before any expected changes in structure, such as left ventricular hypertrophy or carotid atheroma.
There is evidence to suggest an interdependence of the vascular inflammatory process, as reflected by high sensitivity C-reactive protein, and the presence of vasomotor endothelial dysfunction. Although both may also correlate with overall Framingham risk calculations, neither is highly correlated to individual risk factors or when Framingham risk is low (15,16). Thus, a second limitation shared with all other research in this field is a lack of analysis of the predictive role of endothelial dysfunction measurements when high sensitivity C-reactive protein measurements are also available in the same patients.
The investigators should be congratulated for putting endothelial dysfunction assessment into a context that includes many of the standard testing methodologies. But the conclusions should not be extrapolated without due regard to the cohort upon which the conclusions are based. The ultimate value of this work is to encourage further studies of a similarly rigorous and clinically relevant nature to determine when novel risk assessment methods might be of greatest use. In the area of endothelial dysfunction measurement, there is a striking need to determine the prognostic value of the approach in a cohort of patients for whom cardiovascular risk prediction is more tenuous than in most of the populations studied to date. Virtually none of the articles reviewed at this time pertain solely to populations without manifest cardiovascular disease or without diseases that have a high a priori likelihood of underlying vascular disease. Novel risk assessment methods that go beyond the current algorithms, such as the Framingham risk assessment algorithm, may well find their niche in uncomplicated patients at low or moderate risk of events. Even so, endothelial dysfunction measurement must be clearly shown to be applicable in a clinical setting by developing more practical and reproducible methodologies and by standardizing the reporting of results. These milestones have not been achieved yet (17). Therefore, although it is true that endothelial dysfunction testing is not yet ready for "prime time," it is also true that the quest to determine the optimum way to use this approach for assessing prognosis and for monitoring treatment should continue undeterred.
 |
Footnotes |
|---|
|
References |
|---|
 Top References |
|---|
This article has been cited by other articles:
|
|
 |
|
  Y.-H. Chan, K.-K. Lau, K.-H. Yiu, S.-W. Li, H.-T. Chan, D. Y.-T. Fong, S. Tam, C.-P. Lau, and H.-F. Tse Reduction of C-reactive protein with isoflavone supplement reverses endothelial dysfunction in patients with ischaemic stroke Eur. Heart J., September 23, 2008; (2008) ehn409v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Zheng, M. Patel, K. Hryniewicz, and S. D. Katz Association of extended work shifts, vascular function, and inflammatory markers in internal medicine residents: a randomized crossover trial. JAMA, September 6, 2006; 296(9): 1049 - 1050. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Feletou and P. M. Vanhoutte Endothelial dysfunction: a multifaceted disorder (The Wiggers Award Lecture) Am J Physiol Heart Circ Physiol, September 1, 2006; 291(3): H985 - H1002. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. G. Howarth, W. B. Wiehler, M. Pannirselvam, Y. Jiang, J. P. Berger, D. Severson, T. J. Anderson, and C. R. Triggle A Nonthiazolidinedione Peroxisome Proliferator-Activated Receptor {gamma} Agonist Reverses Endothelial Dysfunction in Diabetic (db/db-/-) Mice J. Pharmacol. Exp. Ther., January 1, 2006; 316(1): 364 - 370. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Sherwood, A. L. Hinderliter, L. L. Watkins, R. A. Waugh, and J. A. Blumenthal Impaired Endothelial Function in Coronary Heart Disease Patients With Depressive Symptomatology J. Am. Coll. Cardiol., August 16, 2005; 46(4): 656 - 659. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Zheng, R. Cable, B. Spencer, N. Votto, and S. D. Katz Iron Stores and Vascular Function in Voluntary Blood Donors Arterioscler. Thromb. Vasc. Biol., August 1, 2005; 25(8): 1577 - 1583. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Ding, A. G. Howarth, M. Pannirselvam, T. J. Anderson, D. L. Severson, W. B. Wiehler, C. R. Triggle, and B. S. Tuana Endothelial dysfunction in Type 2 diabetes correlates with deregulated expression of the tail-anchored membrane protein SLMAP Am J Physiol Heart Circ Physiol, July 1, 2005; 289(1): H206 - H211. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. T. Yan, T. J. Anderson, F. Charbonneau, L. Title, S. Verma, E. Lonn, and on behalf of the FATE Investigators Relationship Between Carotid Artery Intima-Media Thickness and Brachial Artery Flow-Mediated Dilation in Middle-Aged Healthy Men J. Am. Coll. Cardiol., June 21, 2005; 45(12): 1980 - 1986. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. D. Katz, K. Hryniewicz, I. Hriljac, K. Balidemaj, C. Dimayuga, A. Hudaihed, and A. Yasskiy Vascular Endothelial Dysfunction and Mortality Risk in Patients With Chronic Heart Failure Circulation, January 25, 2005; 111(3): 310 - 314. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |
