- Email: [email protected]
Fenofibrate and diabetic retinopathy
Correspondence
5
Wong TY, Cheung N, Islam FM, et al. The relationship of retinopathy to coronary artery calcification: the Multi-Ethnic Study of Atherosclerosis. Am J Epidemiol 2008; 167: 51–58.
A C Keech and colleagues1 are to be commended on their monumental FIELD study, but I write to express amazement that they were able to find nearly 10 000 patients with diabetes who were “without requiring lipidmodifying treatment at study entry”. I very rarely see patients with diabetes who are not on a statin, and, given that at entry to the study more than 20% of the participants already had cardiovascular disease, it is little short of incredible that so many patients seem to have been denied standard treatment. My questions for Keech and colleagues are: (1) can they confirm that no patient received a statin at any time during the study? and (2) if they do confirm this, do they think that their findings are relevant to the management of the many patients with diabetes who are receiving a statin? I declare that I have no conflict of interest.
John Firth jfi[email protected] Cambridge University NHS Foundation Trust, Cambridge BC2 8QQ, UK 1
Keech AC, Mitchell P, Summanen P, et al, for the FIELD study investigators. Effect of fenofibrate on the need for laser treatment for diabetic retinopathy (FIELD study): a randomised controlled trial. Lancet 2007; 370: 1687–97.
Authors’ reply
For information on the ACCORDEYE substudy see http://www. accordanc.org/eyeinfo.cfm
722
We accept Gerald Liew and colleagues’ suggestion that statins might yet also prove to have a role in retinopathy prevention, although the studies they reference are preliminary and need confirmation. In the randomised controlled CARDS trial, atorvastatin did not significantly reduce the rate of laser requirement for retinopathy compared with placebo (hazard ratio 0·87, 95% CI 0·65–1·17, p=0·36).1 The ACCORDEYE study will mainly provide evidence on the role of fenofibrate added to a statin, not on the statin itself. Of course, if statins are found to reduce
laser requirement for retinopathy, then the differential statin uptake during follow-up in FIELD (much more so in those allocated placebo than fenofibrate2) would mean that the observed 31% lower rate of laser with fenofibrate would be an underestimate of its real protective effect. As Ning Cheung and T Y Wong point out, our substudy was not ideally powered to examine differences in rates of retinopathy progression, highlighting the challenge of identifying relevant previous cohorts on which to base assumed event rates. The observed placebo-group two-step progression rate of around 12% was less than half of that expected, meaning that the substudy power to detect a 25% reduction in progression was only around 36%. However, lack of power is not an issue concerning the observed positive effects on the reduction in the need for laser therapy, a prespecified tertiary endpoint, seen in both the main trial and the substudy. Also, the observed benefits of fenofibrate in FIELD on a range of cardiovascular events, hospital admission for acute coronary syndromes, amputations, microalbuminuria, and retinopathy2,3 could be consistent with a possible “unifying mechanism” across microvascular and macrovascular outcomes, although FIELD provides no evidence to date that this is the case. John Firth’s comments underscore the importance of designing and conducting clinical trials in the light of knowledge of other concurrent research. On the basis of the wide benefits of statins in diabetes recently published in The Lancet,4 statins should now be considered routinely in most patients with diabetes. But at the time FIELD was started, most patients with diabetes were not considered for statin therapy (and several placebocontrolled trials of statins in diabetes were ongoing through most of the FIELD follow-up period). Both for pragmatic and ethical reasons, FIELD recruited patients who were not yet regarded by their treating doctors (or
national treatment guidelines at the time) as needing statin therapy. After randomisation, however, patients could be started on a statin if regarded as indicated, at the sole discretion of their usual doctors. This was most likely to arise if a patient’s clinical condition changed (such as development of new clinical coronary disease meeting local guidelines for treatment), or if the global evidence about the value of statin therapy in diabetes changed. With the release of the Heart Protection Study findings in 2002,5 FIELD patients and their doctors were notified, and many patients did elect to start statin therapy. The uptake of statins was more common among those allocated to placebo than fenofibrate (around 34% vs 19%, respectively, by study close). This was allowed for in the study design, but was thought to have significantly attenuated the measured intentionto-treat effects of fenofibrate. Nevertheless, this policy always maintained the patients’ best interests foremost. We declare that we have no conflict of interest other than that stated in the original paper.
*A C Keech, P M Mitchell, P A Summanen, T M E Davis, R J Simes [email protected] NHMRC Clinical Trials Centre, University of Sydney, Building F, 88 Mallet Street, Camperdown, New South Wales 2050, Australia 1
2
3
4
5
Bloomgarden ZT. Diabetic Retinopathy and diabetic neuropathy. Diabetes Care 2007; 30: 760–65. Keech A, Simes RJ, Barter P, et al, for the FIELD study investigators. Effects of long-term fenofibrate therapy on cardiovascular events in 9795 people with type 2 diabetes mellitus (the FIELD study): randomised controlled trial. Lancet 2005; 366: 1849–61. Burgess D, Hunt D, Li L, et al. Effects of fenofibrate on silent myocardial infarction, hospitalization for acute coronary syndromes and amputation in type 2 diabetes: the Fenofibrate Intervention and Event Lowering in Diabetes (FIELD) study. Circulation 2007; 116 (suppl II): 838 (abstr 3693). Cholesterol Treatment Trialists’ Collaborators. Efficacy of cholesterol-lowering therapy in 18 686 people with diabetes in 14 randomised trials of statins: a meta-analysis. Lancet 2008; 371: 117–25. Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of cholesterol lowering with simvastatin in 20 536 high-risk individuals: a randomised placebo-controlled trial. Lancet 2002; 360: 7–22.
www.thelancet.com Vol 371 March 1, 2008
5
Wong TY, Cheung N, Islam FM, et al. The relationship of retinopathy to coronary artery calcification: the Multi-Ethnic Study of Atherosclerosis. Am J Epidemiol 2008; 167: 51–58.
A C Keech and colleagues1 are to be commended on their monumental FIELD study, but I write to express amazement that they were able to find nearly 10 000 patients with diabetes who were “without requiring lipidmodifying treatment at study entry”. I very rarely see patients with diabetes who are not on a statin, and, given that at entry to the study more than 20% of the participants already had cardiovascular disease, it is little short of incredible that so many patients seem to have been denied standard treatment. My questions for Keech and colleagues are: (1) can they confirm that no patient received a statin at any time during the study? and (2) if they do confirm this, do they think that their findings are relevant to the management of the many patients with diabetes who are receiving a statin? I declare that I have no conflict of interest.
John Firth jfi[email protected] Cambridge University NHS Foundation Trust, Cambridge BC2 8QQ, UK 1
Keech AC, Mitchell P, Summanen P, et al, for the FIELD study investigators. Effect of fenofibrate on the need for laser treatment for diabetic retinopathy (FIELD study): a randomised controlled trial. Lancet 2007; 370: 1687–97.
Authors’ reply
For information on the ACCORDEYE substudy see http://www. accordanc.org/eyeinfo.cfm
722
We accept Gerald Liew and colleagues’ suggestion that statins might yet also prove to have a role in retinopathy prevention, although the studies they reference are preliminary and need confirmation. In the randomised controlled CARDS trial, atorvastatin did not significantly reduce the rate of laser requirement for retinopathy compared with placebo (hazard ratio 0·87, 95% CI 0·65–1·17, p=0·36).1 The ACCORDEYE study will mainly provide evidence on the role of fenofibrate added to a statin, not on the statin itself. Of course, if statins are found to reduce
laser requirement for retinopathy, then the differential statin uptake during follow-up in FIELD (much more so in those allocated placebo than fenofibrate2) would mean that the observed 31% lower rate of laser with fenofibrate would be an underestimate of its real protective effect. As Ning Cheung and T Y Wong point out, our substudy was not ideally powered to examine differences in rates of retinopathy progression, highlighting the challenge of identifying relevant previous cohorts on which to base assumed event rates. The observed placebo-group two-step progression rate of around 12% was less than half of that expected, meaning that the substudy power to detect a 25% reduction in progression was only around 36%. However, lack of power is not an issue concerning the observed positive effects on the reduction in the need for laser therapy, a prespecified tertiary endpoint, seen in both the main trial and the substudy. Also, the observed benefits of fenofibrate in FIELD on a range of cardiovascular events, hospital admission for acute coronary syndromes, amputations, microalbuminuria, and retinopathy2,3 could be consistent with a possible “unifying mechanism” across microvascular and macrovascular outcomes, although FIELD provides no evidence to date that this is the case. John Firth’s comments underscore the importance of designing and conducting clinical trials in the light of knowledge of other concurrent research. On the basis of the wide benefits of statins in diabetes recently published in The Lancet,4 statins should now be considered routinely in most patients with diabetes. But at the time FIELD was started, most patients with diabetes were not considered for statin therapy (and several placebocontrolled trials of statins in diabetes were ongoing through most of the FIELD follow-up period). Both for pragmatic and ethical reasons, FIELD recruited patients who were not yet regarded by their treating doctors (or
national treatment guidelines at the time) as needing statin therapy. After randomisation, however, patients could be started on a statin if regarded as indicated, at the sole discretion of their usual doctors. This was most likely to arise if a patient’s clinical condition changed (such as development of new clinical coronary disease meeting local guidelines for treatment), or if the global evidence about the value of statin therapy in diabetes changed. With the release of the Heart Protection Study findings in 2002,5 FIELD patients and their doctors were notified, and many patients did elect to start statin therapy. The uptake of statins was more common among those allocated to placebo than fenofibrate (around 34% vs 19%, respectively, by study close). This was allowed for in the study design, but was thought to have significantly attenuated the measured intentionto-treat effects of fenofibrate. Nevertheless, this policy always maintained the patients’ best interests foremost. We declare that we have no conflict of interest other than that stated in the original paper.
*A C Keech, P M Mitchell, P A Summanen, T M E Davis, R J Simes [email protected] NHMRC Clinical Trials Centre, University of Sydney, Building F, 88 Mallet Street, Camperdown, New South Wales 2050, Australia 1
2
3
4
5
Bloomgarden ZT. Diabetic Retinopathy and diabetic neuropathy. Diabetes Care 2007; 30: 760–65. Keech A, Simes RJ, Barter P, et al, for the FIELD study investigators. Effects of long-term fenofibrate therapy on cardiovascular events in 9795 people with type 2 diabetes mellitus (the FIELD study): randomised controlled trial. Lancet 2005; 366: 1849–61. Burgess D, Hunt D, Li L, et al. Effects of fenofibrate on silent myocardial infarction, hospitalization for acute coronary syndromes and amputation in type 2 diabetes: the Fenofibrate Intervention and Event Lowering in Diabetes (FIELD) study. Circulation 2007; 116 (suppl II): 838 (abstr 3693). Cholesterol Treatment Trialists’ Collaborators. Efficacy of cholesterol-lowering therapy in 18 686 people with diabetes in 14 randomised trials of statins: a meta-analysis. Lancet 2008; 371: 117–25. Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of cholesterol lowering with simvastatin in 20 536 high-risk individuals: a randomised placebo-controlled trial. Lancet 2002; 360: 7–22.
www.thelancet.com Vol 371 March 1, 2008