Dhananjay Vaidya, School of Medicine, Baltimore, MD 21287, USA, Assistant Professor of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD 21287, USA

Introduction

The metabolic syndrome was proposed as an entity ~20 years ago [1], however questions about the clinical utility of the concept continue to be raised. Some conclude from published data that the concept is not useful [2,3], while others conclude that the data support its use [4]. This raises the doubt that experts may differ about what constitutes utility, but this issue has not been explicitly discussed by most commentators.
Many recent examinations of the clinical utility of the metabolic syndrome test whether the co-occurrence of metabolic abnormalities is associated with greater pathology, morbidity, and mortality above and beyond the simple addition of separate component associations [5-7]. Recognizing that this is not the only utility of postulating clinical entities, we will apply to the metabolic syndrome insights gleaned from the motivation for three other well-known syndromes: (1) Down’s syndrome, (2) acquired immunodeficiency syndrome, and (3) the Duckett-Jones criteria for rheumatic fever.

An Informal Look at Other “Syndromes”

DOWN’S SYNDROME

In 1866, Down wrote that the number of his “feeble-minded” pediatric patients with “mongoloid” facial features was “so great”, that it constituted 10% of the cases presented to him [8]. Obviously, 10% represents a great number because this was higher than expected in his patient population. The utility according to Down was in expecting certain other mental characteristics relevant for treatment and trainability, anticipating the prognosis of lower life expectancy, and a tendency towards tuberculosis, and also in advising the parents that the cause was congenital [8]. A clinical diagnostic criterion list for the Down’s syndrome was compiled in 1980 [9], when routine karyotyping was not feasible. For this list, there was a saturation of diagnostic efficiency for the chromosomal abnormality at high list component counts [9], which is statistically equivalent to a sub-additive association of component count.

Acquired Immunodeficiency Syndrome

In 1982, Fauci commented on the “syndrome” of Kaposi’s sarcoma and certain opportunistic infections among previously healthy homosexual men [10]. With only 290 cases recognized at that time, the alarm and impulse to identify the entity stemmed from the cluster of clinical diagnoses being “unprecedented” [10]. The chief suspected cause of that time, cytomegalovirus, was later acquitted, and no strong argument was made regarding supra-additive risk. The utility in identifying this clustering as an entity was to set the goal of finding “the mystery behind this extraordinary disease” [10]. Even after the availability of tests for infection by the human immunodeficiency virus in the late 1980s, the diagnosis continued to be made based on a list of clinical criteria [11], which was relatively cheap and convenient in spite of imperfect sensitivity and specificity [12].

Rheumatic Fever

For rheumatic fever, the only morbid or fatal sequelae are related to carditis [13], and not other clinical features of the disease. Though supra-additive morbidity from the various non-carditis components is moot, the list of clinical findings that constitute the Duckett-Jones criteria is useful for diagnosis of rheumatic fever [14]. In fact, the mandate for early treatment of streptococcal infections in suspected rheumatic fever makes it likely that direct microbiological evidence cannot be readily obtained for diagnosis.

Different Motivations Suggest Different Meanings for “Syndrome”

This cursory but broad sweep of other clinical entities suggests alternative motivations for component-lists in the clinic or public health. Regarding the Duckett Jones criteria there was a known unifying cause to be diagnosed by proxy [14]; for the acquired immunodeficiency syndrome, the unusual clustering in 1982 hinted at a unifying but unknown cause [10]; and in addition to a speculative [8] or known [9] unifying cause for the Down’s syndrome, there was also a unifying bad prognosis [8], which may be thought of as supra-additive risk, or diagnostic convenience [9]. Indeed, even for defining the metabolic syndrome, Kim and Reaven [15] identified the proxy diagnosis of insulin resistance and the prediction of cardiovascular risk as disparate goals. Whatever may constitute utility for a given syndrome, it is important that the concept be operationalized, so that “utility” could be quantified using data. In a recent paper [16], we suggested that there were multiple ways to operationalize various concepts, each construct with distinct meaning.

Supra Additive Consequences of Co-Occurrence

One valid construct already introduced is that co-prevalent components should be associated with greater pathology, morbidity of mortality as compared to the additive association of each of the components. Reports of supra-additive association of the metabolic syndrome with subclinical atherosclerosis as measured by carotid intimal-medial thickness are equivocal [7,16], while others have shown no supra-additive association with incident cardiovascular disease [5] or premature coronary disease [6]. Thus no supra-additive mortality benefit is presumably expected if there is partial treatment success and only some of the syndrome components are addressed.

Unexpected Degree of Co-Occurrence

Another construct relates to “unusual co-occurrence.” The straightforward metric for this is to compare the observed and expected co-prevalence by chance of multiple metabolic components, using standard χ2 statistics. If components tend to cluster, higher component counts and zero component counts will be observed more frequently than expected, while intermediate component counts will occur less frequently than expected. We found this to be the case in the Multiethnic Study of Atherosclerosis (MESA) for the NCEP-ATP-III metabolic syndrome, with zero components observed in 12% of men and 11% of women, though only 5% of men and 3% of women should have been free of all components by chance [16]. A co-prevalence of all 5 components was seen in 5% of men and 7% of women, though the expectation by chance was 1% and 2%, respectively [16]. This “unusual co-occurrence” merits the investigation of a common cause, and for this purpose, naming of the “metabolic syndrome” entity would be justified. There is no unanimity whether the root cause is insulin resistance or obesity, and the cause may well be a currently undefined social or behavioral factor.

A Checklist of Diagnostic Criteria for a Known Condition

Yet another construct tries to capture the utility of the metabolic component list as a diagnostic proxy for insulin resistance, if direct measurement is not feasible. An additive association is adequate for this purpose. In MESA, we found that each of the components of the NCEP-ATP-III metabolic syndrome was associated with a higher measure homeostatic model assessment of insulin resistance product (HOMA = fasting glucose x insulin) [16]. Further, a higher component count was associated with greater HOMA levels throughout the range of component counts from none to all five [16]. Though there is no established dichotomous definition for abnormal levels of HOMA, for the heuristic definition of insulin resistance as the upper quartile of HOMA, the area under the Receiver Operating Characteristic curve for the component count was 0.80 in both men and women [16]. This is fairly good diagnostic efficiency for a non-acute condition such as insulin resistance.

An Approach to Open Questions Regarding the “Syndrome”

For all constructs of the meaning of “syndrome,” some critiques of the current definitions remain pertinent. What logic guides the choice of exactly 5 or 6 metabolic abnormalities excluding others? How does one arrive at the thresholds of abnormality for each of the metabolic assays? Should all of the components be given equal weight or should some components be more equal than others? For devising a convenient diagnostic proxy, there may be value in keeping the list short and limiting the weights of each component to simple whole numbers for easy calculation. Certain biochemical or imaging assays may not be enumerated in the list if they are expensive or painful. Such decisions need an expert consensus regarding pragmatics and cannot be purely data driven. Once the pragmatic parameters are set, data may used to derive the most parsimonious component list that achieves a practically acceptable level of false negatives and false positives. On the other hand, in the investigation for an underlying cause or common biochemical pathways, a longer list of components advances biomedical understanding, if it can be demonstrated that many more phenomena are connected by the unifying theme [17]. Even for the two constructs that demonstrate the utility of the metabolic syndrome [16], the decision about utility will guide divergent scientific enquiry towards parsimony or expansion of the component list.

Conclusion

The metabolic syndrome perhaps fails the test of supra-additive risk of morbidity and mortality [5,6], and also of subclinical pathology [16], though there may also be evidence to the contrary [7]. However, in analogy to other clinical conditions, we recognize at least two other valid clinical uses for recognizing the list of metabolic abnormalities as an entity [16]. The debate regarding the validity of the metabolic syndrome needs an introspective pause so that clinical and public health practitioners agree on what constitutes the utility of recognizing the syndrome as a disease entity.

Acknowledgements

I thank Dr. Pamela Ouyang and Dr. Frederick Brancati for critically reading through this commentary.

References

1. Reaven GM. Banting lecture 1988. Role of insulin resistance in human disease. Diabetes 1988;37:1595-607.
2. Kahn R, Buse J, Ferrannini E, Stern M. The metabolic syndrome: time for a critical appraisal: joint statement from the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care 2005;28:2289-304.
3. Kahn R, Buse J, Ferrannini E, Stern M. Author Response to "The metabolic syndrome: perhaps an etiologic mystery but far from a myth -- where does the International Diabetes Federation stand?" Medscape Diabetes and Endocrinology, 2005.
4. Zimmet PZ, Alberti G. The metabolic syndrome: perhaps an etiologic mystery but far from a myth - where does the International Diabetes Federation stand? Medscape Diabetes and Endocrinology, 2005.
5. Sundstrom J, Vallhagen E, Riserus U, et al. Risk associated with the metabolic syndrome versus the sum of its individual components. Diabetes Care 2006;29:1673-74.
6. Iribarren C, Go AS, Husson G, et al. Metabolic syndrome and early-onset coronary artery disease: is the whole greater than its parts? J Am Coll Cardiol 2006;48:1800-7.
7. Golden SH, Folsom AR, Coresh J, Sharrett AR, Szklo M, Brancati F. Risk factor groupings related to insulin resistance and their synergistic effects on subclinical atherosclerosis: the atherosclerosis risk in communities study. Diabetes 2002;51:3069-76.
8. Down JLH. Observations on an ethnic classification of idiots. London Hospital Reports 1866;3:S259-S262.
9. Fried K. A score based on eight signs in the diagnosis of Down syndrome in the newborn. J Ment Defic Res 1980;24:181-85.
10. Fauci AS. The syndrome of Kaposi's sarcoma and opportunistic infections: an epidemiologically restricted disorder of immunoregulation. Ann Intern Med 1982;96:777-79.
11. WHO. Acquired immune deficiency syndrome (AIDS). Revision of the case definition of AIDS. Weekly Epidemiological Record 1985;60:270-71.
12. Keou FX, Belec L, Esunge PM, Cancre N, Gresenguet G. World Health Organization clinical case definition for AIDS in Africa: an analysis of evaluations. East Afr Med J 1992;69:550-53.
13. Olgunturk R, Canter B, Tunaoglu FS, Kula S. Review of 609 patients with rheumatic fever in terms of revised and updated Jones criteria. Int J Cardiol 2006 Sep 10;112(1):91-8. Epub 2005 Dec 20.
14. Guidelines for the diagnosis of rheumatic fever. Jones Criteria, 1992 update. Special Writing Group of the Committee on Rheumatic Fever, Endocarditis, and Kawasaki Disease of the Council on Cardiovascular Disease in the Young of the American Heart Association. JAMA 1992;268:2069-73.
15. Kim SH, Reaven GM. The metabolic syndrome: one step forward, two steps back. Diab Vasc Dis Res 2004;1:68-75.
16. Vaidya D, Szklo M, Liu K, Schreiner PJ, Bertoni AG, Ouyang P. Defining the metabolic syndrome construct: Multi-Ethnic Study of Atherosclerosis Cross-sectional Analysis. Diabetes Care 2007 May 7; [Epub ahead of print].
17. Reaven GM. Why Syndrome X? From Harold Himsworth to the insulin resistance syndrome. Cell Metab 2005;1:9-14.

Please address correspondence to:
Dhananjay Vaidya, Ph.D., M.P.H.
1830 East Monument St, Room 8028-A
Johns Hopkins Medical Institutions
Baltimore, MD. 21287
Tel: 410-955-7781/410-550-7479
Fax: 410-955-0321
Email: dvaidya1[at]jhmi.edu