The legacy effect in type 2 diabetes: fact or fiction?

Michel MARRE,MD, PhD
Service d’Endocrinologie-Diabète-Nutrition, Groupe Hospitalier Bichat – Claude Bernard Assistance Publique des Hôpitaux de Paris, and INSERM U695, Université
Denis Diderot – Paris 7 – FRANCE

The legacy effect in type 2 diabetes: fact or fiction?

by M. Marre, France

Diabetes is a lifelong disease; thus, it is of utmost importance that we understand the “legacy effect” of long-term interventions on vascular outcomes. Intervention studies conducted many years ago on the diabetic retinal disease in type 1 diabetes gave rise to the concept of “memory effect” in uncontrolled diabetes. A long-term effect of high glucose levels on the functions and structure of vessels, with an impact on outcomes, is biologically plausible. In particular, interventions for glycemic control may lead, in the long term, to a benefit in terms of risk for premature death, as shown in the UKPDS (United Kingdom Prospective Diabetes Study). Interestingly, this legacy effect does not apply to high glucose only, but also to other vascular risk factors, such as hypertension and lipids. As many factors can influence final outcomes over a lifetime, it is important to set up prospective follow-up studies of participants in large vascular clinical trials. However, the duration of follow-up must be long enough to be clinically meaningful. Such considerations lead to implementation of the ADVANCE-ON study (Action in Diabetes and Vascular disease: PreterAx and DiamicroN MR Controlled Evaluation post-trial ObservatioNal study), expected to provide important follow-up data in type 2 diabetes patients in a contemporary diabetes management setting.

Medicographia. 2013;35:53-60 (see French abstract on page 60)

The concept of a glucose “legacy effect” (initially called “memory effect”) emerged from the first trial on the effect of strict glycemic control in retinopathy in patients with insulin-dependent diabetes and background retinopathy.1,2 In this single-center, open-label, randomized trial, participants allocated to continuous subcutaneous insulin infusion had poorer retinal status than those on conventional insulin treatment over a six-month period, although they had improved albumin excretion rates (AER).1 This intergroup difference vanished, however, after one year.2 Similar findings were later observed in the DCCT (Diabetes Control and Complications Trial) among participants with background retinopathy at baseline, during the first two years of the trial.3 From these data came the concept that the body could essentially “remember” long periods with high glucose, and this was called the “memory effect.” This term was later substituted by “legacy effect” when the results of the ten-year follow-up of the UKPDS study were issued.4,5

Since then, questions have arisen with regard to the legacy effect of glucose on vascular outcomes in clinical diabetes. Is the legacy effect permanent or reversible? Would late intervention to control high glucose have an effect on outcomes, especially once diabetes has continued uncontrolled long enough to cause deterioration of vessels? Does the legacy effect apply to some complications (retinal), but not to others (renal)? Is there a biological precedent for such a phenomenon? Is this phenomenon specific to glucose, or is it observable with other risk factors? What are the practical issues for the different treatment strategies? Here, let’s take a closer look at the legacy effect and some of these questions.

Figure 1
Figure 1. Impact of strict glycemic control in the participants of the DCCT.

A sustained change in the severity of retinopathy was defined as a change observed by fundus photography of at least three steps from baseline that was sustained for at least six months. A, Primary-prevention cohort. B, Secondary-intervention cohort. The numbers of patients evaluated in each therapy group are indicated below the graphs. Note the initially negative impact of strict glycemic control on the course of retinopathy in the participants with baseline background retinopathy (Panel B).
Abbreviation: DCCT, Diabetes Control and Complications Trial.
After reference 3: Diabetes Control and Complications Trial Research Group. N Engl J Med. 1993;329(14):977-986. © 1993, Massachusetts Medical Society.

Summary of legacy effects observed in diabetes

Legacy effects were first observed in type 1, insulin-dependent diabetes during the DCCT and its post-trial follow-up, EDIC (Epidemiology of Diabetes Interventions and Complications). These observations illustrated that glucose memory may have both negative and positive effects.

As an illustration of a negative effect, let’s look once again to the principal data of the DCCT.3 The participants were randomized into intensive versus conventional treatment with a factorial design; half of them had no retinopathy at baseline, and half of them had background retinopathy. In those with background retinopathy who were allocated to the intensive arm, retinal status deteriorated slightly up to the third year (although this deterioration was not severe enough to warrant termination of the trial). Afterwards, this memory effect vanished; however, consequently, the final benefit (relative risk reduction) over a 6.5-year period was lower than in those participants without retinopathy at baseline (Figure 1).3

In contrast, most of the follow-up studies published thus far on the DCCT-EDIC study illustrate well a positive effect from memory of a long period with strict glycemic control. First, the carotid intima-media thickness of participants allocated to intensive treatment progressed less than in the control group several years after the end of the study, while mean HbA1c had regressed to an identical mean value for the two groups.6 Second, the intensive group displayed significantly fewer cardiovascular events during follow-up than the control group, with an impressive risk reduction of around 50%.7 Notably, a 41% risk reduction in cardiovascular events was detectable by the end of the DCCT, although numbers were not high enough to reach statistical significance.3 Third, the risk for impaired glomerular filtration rate in the intensive treatment group was reduced by 50% in the long term, compared with the control group.8 Finally, the above-mentioned benefits in the long term were not counterbalanced by any impairment in the cognitive functions of the participants, which could have resulted from a 3-fold higher number of severe hypoglycemic episodes.9

Figure 2
Figure 2. The legacy effect of intensified glycemic control in the UKPDS.

Hazard ratios for patients in the UKPDS who had any diabetes-related end point (A and B), myocardial infarction (C and D), or microvascular disease (E and F) or who died from any cause (G and H) are shown for the sulfonylurea-insulin group versus the conventional-therapy group and for the metformin group versus the conventional-therapy group. The overall values at the end of the study, in 1997, are shown (red squares), along with the annual values during the 10-year posttrial
monitoring period (blue diamonds). Hazard ratios below unity indicate a favorable outcome from sulfonylurea or metformin therapy. Numbers of first events in an
aggregate outcome that accumulated in each group are shown at 2-year intervals. The vertical bars represent 95% confidence intervals.
Abbreviation: UKPDS, United Kingdom Prospective Diabetes Study.
After reference 4: Holman et al. N Engl J Med. 2008;359(15):1577-1589. © 2008, Massachusetts Medical Society.

In type 2 diabetes, all data available as regards legacy effect were generated from the UKPDS study.4,5 In the glucose arm of this open-label, multicenter, national study, the 0.9% HbA1c difference between the intensive treatment (based on the use of sulfonylureas or insulin) and the conventional one (dietary intervention only) produced a significant 25% risk reduction for microvascular outcomes, but no significant effect on risk for myocardial infarction, stroke, or death from any cause.10 Ten years after the study’s end, this group reported the same outcomes for the 10-year follow-up.410 were sustained or improved in the follow-up period, and the long follow-up with much higher number of events revealed substantial and statistically significant benefits. The most important benefit concerned the rate of death from any cause (Figure 2, page 55).4 Other wise, most intergroup hazard ratios remained stable throughout the observation period.

Another study in type 2 diabetics, ACCORD (Action to Control CardiOvascular Risk in Diabetes), was terminated prematurely in February 2008, because the rate of total mortality (including cardiovascular mortality) was higher in the experimental group than in the control group.11,12 Parenthetically, this result on the impact of intensive glycemic control in type 2 diabetes was not confirmed by the meta-analysis of this outcome in all individual participants of the four major trials in this domain: UKPDS, ADVANCE, ACCORD, and VADT (Veterans Affairs Diabetes Trial).13 Because the ACCORD participants were also randomized to lipid-lowering, and to blood pressure–lowering studies, they were followed-up accordingly, in ACCORD-ON, for the outcome of the glucose arm of the study until June 2009, the originally planned end.11 Data collected during the 16-month extra follow-up were similar to the initially published data. However, the follow-up period was probably too short to eliminate the concerns that led to the early termination of the ACCORD trial, from a methodological viewpoint.

Legacy effect: permanent and applicable to all diabetes-related outcomes?

As mentioned earlier, a negative memory effect on retinal condition was reported to vanish in participants of the first trial for which this phenomenon was reported.1,2 Moreover, the effect on AER was beneficial from the first six months of the study.1 However, the AER is a functional parameter, not an anatomical one, nor an outcome per se.

Figure 3
Figure 3. Course of glomerular lesions
in a patient cured from insulindependent
diabetes over 10 years.

A, A typical glomerulus from the baseline
biopsy specimen, which is characterized
by diffuse and nodular (Kimmelstiel-Wilson)
diabetic glomerulopathy. Mesangial-matrix
expansion and the palisading of mesangial
nuclei around the nodular lesions are evident.
B, A typical glomerulus five years after transplantation
shows the persistence of the diffuse
and nodular lesions. C, A typical glomerulus
10 years after transplantation, with marked
resolution of diffuse and nodular mesangial
lesions and more open glomerular capillary
After reference 16: Fioretto et al. N Engl J
Med. 1998;339(2):69-75. © 1998, Massachusetts
Medical Society.

This issue of functional versus anatomical definitions of complications is well illustrated by the studies conducted by Mauer et al: in 1975, Mauer reported that the anatomical signs of diabetic nephropathy (increased width of the glomerular basement membrane [GBM], and expansion of the mesangium) observed in rats made diabetic with streptozotocin could be reversed once diabetes was cured with islet transplantation.14 To verify if these findings could be translated to human diabetic patients, he carried out a prospective study on insulindependent diabetic patients in Minneapolis who underwent successful pancreas transplantations to cure their diabetes. At baseline, renal biopsies from those without significant renal impairment showed lesions typical of diabetic glomerulopathy. The lesions in the second biopsies performed 5 years later were not much different.15 However, 8 subjects underwent a third renal biopsy 10 years after they were cured of diabetes by pancreas transplantation. These biopsies showed very significant reduction in all the pathological signs of glomerular disease, and AER in these patients was satisfactory (Figure 3).16

Along this line, the data on the effect of strict glycemic control in type 2 diabetic patients with high cardiovascular risk in the ADVANCE study should be kept in mind: not only did this glycemic control prevent microalbuminuria and macroalbuminuria, and perhaps protect against end-stage kidney dis- ease, it also allowed reversal of already established microalbuminuria and macroalbuminuria, perhaps protecting against future premature death and vascular outcomes predictable from baseline AER (Figure 4).17,18 Moreover, similar findings on the effect of strict glycemic control were reported in the ACCORD12 and VADT19 studies.

Thus, high-glucose memory effects on functional and anatomical signs of diabetic vascular disease can be reversed over time.

Figure 4
Figure 4. Impact of updated albumin excretion rates and of glomerular filtration rates on risk for adverse vascular outcomes during the
ADVANCE study.

Abbreviations: ADVANCE, Action in Diabetes and Vascular disease: PreterAx and DiamicroN MR Controlled Evaluation; CI, confidence interval; eGFR, estimated glomerular filtration rate; HR, hazard ratio; SBP, systolic blood pressure; UACR, urinary albumin-to-creatinine ratio.
After reference 17: Ninomiya et al. J Am Soc Nephrol. 2009;20(8):1813-1821. © 2009, American Society of Nephrology.

Legacy effect: observable for variables other than glucose?

Posttrial effects of intervention to control particular risk factors, including abnormal lipid profiles and blood pressure, have been studied in the cardiovascular domain. For example, as WOSCOP (West Of Scotland COronary Prevention study) initially showed that primary intervention with pravastatin 40 mg per day was beneficial (compared with placebo) in terms of risk for cardiovascular death, myocardial infarction, and stroke,20 a follow-up study was implemented to examine continuation of the benefit over several years after trial cessation. Indeed, the hazard ratio was almost unchanged between groups, thereby increasing the initially observed benefit, although blood lipids regressed to similar levels in the active drug and in the placebo groups at trial’s end.21

Similar findings were reported following the HOPE study (Heart Outcomes Prevention Evaluation): 2.6 years after cessation of exposure to ramipril 10 mg per day or to placebo, benefits on risk for myocardial infarction or diabetes onset were still observable, with reductions in relative risks of 19% and 34%, respectively.22 These hazard ratios were not different from those observed in the HOPE trial.23 The follow-up period was relatively short, and the hazard ratios did not change over time.

On the other hand, the follow-up of the UKPDS blood pressure arm was disappointing in that no legacy effect was found from the use of intensified blood pressure treatment.5 A reduced risk for microvascular disease persisted for only 3-4 years, but all other risks were similar for the arms after 1-2 years.

Thus it seems from these studies that a memory (or legacy) effect is more easily detectable for interventions aiming to control blood glucose than those targeting blood lipids or blood pressure. However, several factors seem important: first, the nature of the studied variable; second, the nature of the studied outcome; and third, the duration of observation. Regarding nature of the studied variable, let’s look at blood pressure, for example. This is a hemodynamic variable that can have an impact in the short term (as illustrated by the brilliant and, at that time, unexpected effects of intensified blood pres- sure lowering on microvascular outcomes in the UKPDS24), but whose benefit can vanish early after intervention cessation. This should be kept in mind when considering the HOPE-TOO study (HOPE-The Ongoing Outcomes), whose duration was relatively (2.5 years) short, compared with that of the UKPDS follow-up study. As regards nature of the studied outcome, one should consider that the legacy effect may be due to alterations in anatomy, not only alteration in functional outcomes. This is illustrated by the data on renal outcomes in type 1 diabetic patients. Regarding the duration of observation, as described above, several years of strict glycemic control may be necessary to allow regression of anatomical damage (and reduction of GBM width or mesangial expansion in remnant glomeruli does not mean revival of those occluded by sclerosis), while some weeks or days of reduction in high blood pressure suffice to reduce AER. Thus, the longer the duration of the initial intervention, the more plausible a legacy effect; and the longer the posttrial observation, the more reliable the posttrial data.

Figure 5
Figure 5. Role of mitochondrial alterations
in “glucose memory.”

Abbreviations: AGE, advanced glycation end product;
mtDNA, mitochondrial DNA; RAGE, receptor for advanced
glycation end products; ROS, reactive oxygen species.
After reference 29: Ceriello et al. J Clin Endocrinol Metab.
2009;94(2):410-415. © 2009, The Endocrine Society.

It is highly plausible that intervention against high glucose affects both the biochemical and the hemodynamic components of vascular disease, referring to now old hypotheses both on the biochemical and hemodynamic origins of diabetic microangiopathy.25,26 Both these components can have very positive interactions.27

Possible mechanisms behind the observed effects

Soon after the first reports of a possible negative effect of short-term transition from uncontrolled hyperglycemia to strict glycemic control in patients with insulin-dependent diabetes and background retinopathy,1 an experimental model was set up to examine this phenomenon.28 The studies were conducted on the retinas of dogs rendered diabetic with all oxan and followed over a 5-year period.

The first group was left with uncontrolled diabetes, and retinas displayed typical, severe, diabetic retinopathy at study’s end. The second group was treated intensively for diabetes, and it displayed no, or minimal diabetic retinopathy. The third group was left uncontrolled (as was the first one) for the first 2.5 years, and then treated as was the second group; at study end, the retinas were almost as severely damaged as in the first group.29 These data illustrated well a role for primary prevention of diabetic complications by strict glycemic control from the earliest time after diabetes diagnosis, and the difficulty to reverse established lesions, even those at early stages.

However, the biochemical mechanisms behind the memory phenomenon remained to be studied. A comprehensive review of the literature was produced by Ceriello et al.29 The first biochemical data were reported on the composition of the basement membrane: components such as collagen (hyperglycemia favoring production of collagen IV and fibronectin) which may make it more permeable with functional changes that favor establishment of microvascular disease.30 Furthermore, changes in glycation may alter the electrical charges of proteins, contributing further to alterations in their properties and/or functions. Half-life of these components may be especially long, and this may contribute to the formation of lesions, accounting in part for the glucose memory phenomenon. Many proteins are sensitive to glycation (eg, hemoglobin), and as the half-life of such products is considerably variable, they contribute to functional and structural consequences in many organs in the long term.29 Indeed, the role of advanced glycation end products (AGEs) in the constitution and permanence of microvascular diseases has been studied in depth.31

Brownlee and his group32 demonstrated the biochemical links between hyperglycemia resulting in excessive production of superoxide anions (O2 -) and the development of diabetic complications through several pathways: increased polyol pathway flux, increased AGE formation, activation of protein kinase C, and increased hexosamine pathway flux. The main source of the increased production of O2 – is the mitochondria, as illustrated in Figure 5.29 How can these O2 – produce long-term effects when their main characteristic is a very short (usually less than one minute) half-life? Probably, they interact with compounds like proteins, lipids, and nucleic acids, modifying their conformations. By doing so, they may affect cellular functions and various organs over a long time. In particular, gene transcription, gene expression, and cellular functions may be modified through changes in mRNAs. Furthermore, gene expression depends on how the genome is arranged, which depends on how genomic DNA is wired. This depends on several mechanisms: histone acetylation/deacetylation, methylation/demethylation, ie, the molecular bases of epigenetics.

The impact of moderate hyperglycemia on structure and function of the organism is illustrated by the fate of offspring of diabetic mothers. Increased risk for renal diseases in offspring of type 2 diabetic mothers that had diabetes before pregnancy versus after pregnancy has been shown in Pima Indians.33 However, all aspects linked to insulin resistance, including genetic bases for hypertension, may have confounded the effect of hyperglycemia per se. To circumvent this issue, we studied kidney function in offspring of type 1 diabetic mothers, with offspring from type 1 diabetic fathers used as controls. We found that offspring of type 1 diabetic mothers had reduced renal functional reserve, a phenomenon compatible with a reduced number of nephrons in the offspring of type 1 diabetic mothers.34

Recently obtained preliminary data suggest that the methylation profile of the genome in offspring of type 1 diabetic mothers is different from that in offspring of type 1 diabetic fathers.35 This finding supports the idea that exposure to moderate hyperglycemia can alter function and perhaps structure in the very long term, and that periods of hyperglycemia affect the genome, leaving an imprint on the future of organisms, especially regarding vasculature.

Perspectives and conclusion

Just how periods of hyperglycemia impact on the vascular fate of diabetic subjects remains to be elucidated. Mechanisms may include hemodynamic anomalies or be more biochemical in nature. Interestingly, the pathways through which they affect vessel structure in the long term may be very similar: the processes of inflammation and, later on, sclerosis that are activated by diabetes and hypertension are the same and include mitogen-activated protein (MAP) kinase and nuclear factor κB (NF-κB) pathways. Thus, the duration of exposure is of utmost importance. Secondary interventions are probably less useful than primary interventions, though any given studied outcome may by nature be more sensitive or less sensitive to such interventions. Intuitively, alterations in structure, such as nephron loss, seem more or less irreversible. It is therefore highly recommended to set up long-term follow-up studies of outcome trials like DCCT, UKPDS, ACCORD, or ADVANCE.

To this end, the prospective ADVANCE-ON study was established to follow-up the participants of ADVANCE. Important data are expected to come from this study, and they are needed as thus far the only available follow-up study in type 2 diabetes is the UKPDS. Unfortunately, the UKPDS is out of step with the contemporary setting of diabetes care (several of the drugs tested then are no longer available, or are not used frequently). Also, today, absolute vascular risk is much lower than it was for people with type 2 diabetes in the 1980s. The UKPDS was a primary intervention study, conducted in patients diagnosed with type 2 diabetes. Thus, data on secondary interventions against hyperglycemia are greatly needed, and ADVANCE-ON is perfectly sized for this purpose. Other contemporary posttrial studies are currently underway, in particular ACCORD-ON (see above), and ORIGINALE (ORIGIN And Legacy Effects), the follow-up of the recently published ORIGIN study (Outcome Reduction with an Initial Glargine INtervention).36 It is important to properly document the long-term impact of various strategies in diabetes, since it is a lifelong condition, and to document the efficacy and safety of the drugs used in these settings. _

1. Steno Study Group. Effect of 6 months of strict metabolic control on eye and kidney function in insulin-dependent diabetics with background retinopathy. Lancet. 1982;1(8264):121-124.
2. Lauritzen T, Frost-Larsen K, Larsen HW, Deckert T. Effect of 1 year of near-normal blood glucose levels on retinopathy in insulin-dependent diabetics. Lancet. 1983;1(8318):200-204.
3. The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med. 1993;329 (14):977-986.
4. Holman RR, Paul SK, Bethel MA, Matthews DR, Neil HA. 10-year follow-up of intensive glucose control in type 2 diabetes. N Engl J Med. 2008;359(15):1577- 1589.
5. Holman RR, Paul SK, Bethel MA, Neil HA, Matthews DR. Long-term follow-up after tight control of blood pressure in type 2 diabetes. N Engl J Med. 2008; 359(15):1565-1576.
6. The Diabetes Control and Complications Trial / Epidemiology of Diabetes Interventions and Complications Research Group. Intensive diabetes therapy and carotid intima-media thickness in type 1 diabetes mellitus. N Engl J Med. 2003; 348(23):2294-2303.
7. The Diabetes Control and Complications Trial / Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) Study Research Group. Intensive diabetes treatment and cardiovascular disease in patients with type 1 diabetes. N Engl J Med. 2005;353:2643-2653.
8. The DCCT/EDIC Research Group. Intensive diabetes therapy and glomerular filtration rate in type 1 diabetes. N Engl J Med. 2011;365:2366-2376.
9. The Diabetes Control and Complications Trial / Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) Study Research Group. Long-term effect of diabetes and its treatment on cognitive function. N Engl J Med. 2007; 356:1842-1852.
10. UK Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet. 1998; 352(9131):837-853.
11. The ACCORD Study Group. Long-term effect of intensive glucose lowering on cardiovascular outcomes. N Engl J Med. 2011;364:818-828.
12. The Action to Control Cardiovascular Risk in Diabetes Study Group. Effects of intensive glucose lowering in type 2 diabetes. N Engl J Med. 2008;358:2545-2559.
13. Control Group, Turnbull FM, Abraira C, et al. Intensive glucose control and macrovascular outcomes in type 2 diabetes. Diabetologia. 2009;52(11):2288- 2298.
14. Mauer SM, Steffes MW, Michael AF, Brown DM. Studies of diabetic nephropathy in animals and man. Diabetes. 1976;25(suppl 2):850-857.
15. Bilous RW, Mauer SM, Sutherland DE, Najarian JS, Goetz FC, Steffes MW. The effects of pancreas transplantation on the glomerular structure of renal allografts in patients with insulin-dependent diabetes. N Engl J Med. 1989;321(2):80-85.
16. Fioretto P, Steffes MW, Sutherland DE, Goetz FC, Mauer M. Reversal of lesions of diabetic nephropathy after pancreas transplantation. N Engl J Med. 1998; 339(2):69-75.
17. Ninomiya T, Perkovic V, de Galan BE, et al; ADVANCE Collaborative Group. Albuminuria and kidney function independently predict cardiovascular and renal outcomes in diabetes. J Am Soc Nephrol. 2009;20(8):1813-1821.
18. ADVANCE Collaborative Group; Patel A, MacMahon S, Chalmers J, et al. Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes. N Engl J Med. 2008;358(24):2560-2572.
19. DuckworthW, Abraira C,Moritz T, et al; VADT Investigators. Glucose control and vascular complications in veterans with type 2 diabetes. N Engl J Med. 2009; 360(2):129-139.
20. Shepherd J, Cobbe SM, Ford I, et al. Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia.West of Scotland Coronary Prevention Study Group. N Engl J Med. 1995;333(20):1301-1307.
21. Ford I, Murray H, Packard CJ, Shepherd J, Macfarlane PW, Cobbe SM; West of Scotland Coronary Prevention Study Group. Long-term follow-up of theWest of Scotland Coronary Prevention Study. N Engl J Med. 2007;357(15):1477-1486.
22. HOPE/HOPE-TOO Study Investigators. Long-term effects of ramipril on cardiovascular events and on diabetes. Results of the HOPE study extension. Circulation. 2005;112:1339-1346.
23. The Heart Outcomes Prevention Evaluation Study Investigators. Effects of an angiotensin-converting-enzyme inhibitor, ramipril, on cardiovascular events in high-risk patients. N Engl J Med. 2000;342(3):145-153.
24. UK Prospective Diabetes Study Group. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. BMJ. 1998;317(7160):703-713.
25. Spiro RG. Search for a biochemical basis of diabetic microangiopathy. Diabetologia. 1976;12(1):1-14.
26. Parving HH, Viberti GC, Keen H, Christiansen JS, Lassen NA. Hemodynamic factors in the genesis of diabeticmicroangiopathy. Metabolism. 1983;32(9):943-949.
27. Cooper ME. Interaction of metabolic and haemodynamic factors in mediating experimental diabetic nephropathy. Diabetologia. 2001;44(11):1957-1972.
28. Engerman RL, Kern TS. Progression of incipient diabetic retinopathy during good glycemic control. Diabetes. 1987;36(7):808-812.
29. Ceriello A, Ihnat MA, Thorpe JE. The “metabolic memory”: is more than just tight glucose control necessary to prevent diabetic complications? J Clin Endocrinol Metab. 2009;94(2):410-415.
30. Roy S, Sala R, Cagliero E, Lorenzi M. Overexpression of fibronectin induced by diabetes or high glucose: phenomenon with a memory. Proc Natl Acad Sci U S A. 1990;87(1):404-408.
31. Brownlee M. Banting Lecture 2004. The pathobiology of diabetic complications: A unifying mechanism. Diabetes. 2005;54:1615-1625.
32. Nishikawa T, Edelstein D, Du XL, et al. Normalizing mitochondrial superoxide production blocks three pathways of hyperglycaemic damage. Nature. 2000; 404(6779):787-790.
33. Nelson RG, Morgenstern H, Bennett PH. Intrauterine diabetes exposure and the risk of renal disease in diabetic Pima Indians. Diabetes. 1998;47(9):1489-1493.
34. Abi Khalil C, Travert F, Fetita S, et al. Fetal exposure to maternal type 1 diabetes is associated with renal dysfunction at adult age. Diabetes. 2010;59(10):2631- 2636.
35. Gautier JF, Porcher R, Fetita LS, et al. Kidney dysfunction in adult offspring exposed in utero to type 1 diabetes is associated with alterations in genome-wide DNA methylation profile. Paper presented at: 48th European Association for the Study of Diabetes Annual Meeting; October 2012; Berlin, Germany.
36. ORIGIN Trial Investigators; Gerstein HC, Bosch J, Dagenais GR, et al. Basal insulin and cardiovascular and other outcomes in dysglycemia. N Engl J Med. 2012;367(4):319-328.

Keywords: advanced glycation end product; epigenetics; hyperglycemia; legacy effect; memory effect; retinopathy; type 2 diabetes