Abstract
The study by Gaya et al. 1 that appears in this issue is important for two quite different reasons. The first is for what it adds to our knowledge about atherogenic dyslipoproteinaemia in patients with type 2 diabetes mellitus and the second for what we can learn about why it has been so difficult for apolipoprotein B (apoB) and apoA-I to be incorporated into clinical practice. As for the first, diabetes undoubtedly multiplies the risk of macrovascular disease and dyslipoproteinaemia is undoubtedly a major mediator of the macrovascular disease in diabetes. Through the prism of lipids, diabetic dyslipidaemia is characterized by hypertriglyceridaemia, normal LDL C and low HDL C – that is, by elevated VLDL, normal LDL and low HDL. This combination of findings made up the argument for fibrates as therapy for these subjects, an argument that was widely accepted, but, nonetheless, fell short when tested in a clinical trial. 2
However, the failure of fibrates to reduce cardiac events in patients with type 2 diabetes is not surprising once we turn our attention from lipids to lipoprotein particles as this study does. An elevated apoB means that LDL particle number is elevated, 3 and this means that abnormalities in the number and composition of LDL particles are a major feature of diabetic dyslipoproteinaemia, a critical finding that was missed when only lipids were considered. 4 The underestimate of atherogenic particle number by LDL C when small dense cholesterol-depleted LDL particles predominate is the biological reason as to why apoB outperforms LDL C as a marker of risk and a measure of the adequacy of LDL-lowering therapy. 3 Fibrates only modestly reduce apoB and therefore only modestly reduce LDL particle number. Statins markedly reduce apoB because they markedly reduce LDL particle number. That is why statins reduce cardiovascular events in patients with type 2 diabetes so effectively whereas fibrates do not.
Among the strengths of the present study are the large sample size and the detailed characterization and comparison of the different groups of diabetics. ApoB is clearly higher, on average, in patients with type 2 diabetes than in normals and higher also, on average, in patients with type 2 diabetes than in patients with type 1 diabetes, another clue, perhaps, that fatty acids are more important determinants of apoB than glucose. 5 However in patients with type 1 diabetes, the concentrations of apoB are not low by any means and, over time, particularly if the diabetic state affects the likelihood that LDL particles that enter the arterial wall will be retained there, the high long-term vascular risk in these patients is not surprising. In this regard, the previous evidence that apoB is the most important lipoprotein-related predictor of vascular risk in patients with type 1 diabetes is worth remembering. 6 Of importance, this study extends the evidence first obtained in Caucasians with type 2 diabetes mellitus 7 to South Asians, the group of us at highest risk of diabetes and vascular disease.
But will this evidence added to all the other evidence that apoB is a better marker than LDL C of the risk of vascular disease change clinical practice? The outcome is far from certain. Clinical practice is governed by guidelines and apoB has yet to enter the guidelines of any country but Canada. 8 No doubt the guideline process has contributed much of value to clinical practice; nevertheless, it is demonstrably imperfect, a reality that has drawn increasing comment. 9–12 Much of this has centred on financial conflicts of interest. Arguably more important, however, are the intellectual conflicts of interest. Most of the members of the guideline panels are experts and have gained that status because they have reached conclusions about the issues in front of them. Most of the experts on lipid panels have been cholesterol-advocates and, understandably, may well find it difficult to reverse longstanding views.
Indeed, I would wager that virtually all of us would have a bias in favour of those ideas we have worked through and adopted. Does this mean that experts by virtue of their expertise are biased and should only serve, if at all, in diminished roles on guideline groups? I think not. Scientists are defined – and define themselves – by the tentative nature of their judgements. We must submit our hypotheses, our individual conceptions of how the world outside us is organized and regulated, to the tests of them we call experiments, the results of which govern whether we can retain or must reframe these hypotheses. We also commit ourselves to engage with the hypotheses of others and to attempt to accommodate the differences between their visions of the external world and ours.
To the extent possible, I think we must follow these same principles and practices in the way we conduct our consensus groups. Thus, the consensus process is a scientific process only if there has been a free and equal contest among the different views of reality. Presentation and reply, point and counter-point, and argument and analysis are the tools we should apply to develop our clinical guidelines. Even long-held conclusions must be justified by data and alternative explanations must be considered seriously. Moreover, whatever conclusions emerge must be recognized as tentative in the same sense and for the same reasons that we accept that all our scientific conclusions are tentative – apparently valid but always subject to future challenge, and almost certain to be radically altered with further improvements in our technology and understanding.
In 2005, the Joint British Guidelines, in a single sentence, dismissed apoB as adding nothing of significance to the standard risk factors. 13 With respect, I disagree and I imagine the authors of this paper would also. Surely, uncovering the importance of LDL in diabetic dyslipoproteinaemia, as the present paper demonstrates, would qualify as significant. Indeed, our failure up to now to appreciate the true role of LDL in diabetic dyslipoproteinaemia and to educate physicians accordingly may account for the fact that sales of fibrates in the United States this year are said to amount to approximately 1.9 billion dollars (US). In any event, the evidence that apoB is superior to LDL C as a marker of the risk of vascular disease and an index of the adequacy of LDL-lowering therapy is now so extensive, so one-sided, and covers so many peoples and so many clinical situations that it can fairly be categorized as decisive. 14 The International Study of Infarct Survival is the most recent – and also one of the most elegantly analysed – to demonstrate this. 15 The gap that counts most now is that between the evidence and the guidelines. The reality is that coronary disease is the major cause of death worldwide and LDL is the most treatable major cause. The human cost of not improving the accuracy with which we measure LDL is real. Until influential guidelines recognize this fact, the individuals that could have been saved will continue to be lost.