Editorial

Preventing Diabetic Retinopathy Progression

The recent report from the Action to Control Cardiovas- treatment assignment persisted, although the HbA1c levels
cular Risk in Diabetes Follow-on (ACCORDION) equalized over the first year after the trial was completed,
Research Group has important implications for ophthal- similar to the findings in the DCCT and EDIC.4
1
mologists observing patients with diabetes. The The long-term beneficial results of intensive therapy in
ACCORDION study has shown once again that intensive type 2 diabetes have been demonstrated again in the recent
glycemic control has long-lasting effects in reducing the publication from the ACCORDION research group.1 The
risk of retinopathy progression. Although the role of ACCORDION study resembled the long-term follow-up of
achieving so-called tight blood glucose control was the DCCT and UKPDS as it carried out an observational
controversial many decades ago, there has been consensus follow-up of the Action to Control Cardiovascular Risk in
on its importance for more than 2 decades. So why is this Diabetes (ACCORD) study cohort, after the randomized
recent article important? treatment allocation ended. With a double 2
2 factorial
The Diabetes Control and Complications Trial (DCCT) design, ACCORD evaluated intensive therapy compared
published the results of its 10-year clinical trial in 1993. The with standard therapy for the control of glycemia and blood
DCCT studied 1441 patients with type 1 diabetes, randomly pressure and evaluated the effect of fenofibrate compared
assigned either to conventional or intensive therapy that with placebo in persons with type 2 diabetes with estab-
achieved glycated hemoglobin (HbA1c) levels of approxi- lished cardiovascular disease (CVD) or who at high risk for
mately 9% versus 7%, respectively. The DCCT conclusively CVD.7 In the glycemia control arm of the ACCORD study,
demonstrated the importance of participants started with a mean
intensive blood glucose control The ACCORDION study has shown once HbA1c of approximately 8.1%
in slowing the microvascular again that intensive glycemic control and were assigned randomly to
complications of diabetes, receive either intensive glycemic
including retinopathy, nephropa-
has long-lasting effects in reducing the control (target HbA1c level,
thy, and neuropathy.2 Nearly all risk of retinopathy progression. <6.0%) or standard control
of the DCCT participants were (target HbA1c level, 7.0%e
followed up subsequently in the observational 7.9%). During the trial, the HbA1c levels decreased to a
Epidemiology of Diabetes Interventions and mean of 6.4% in the intensive group and to 7.5% in the
Complications (EDIC) study.3 During the EDIC study, the standard group. After a mean 3.5 years of follow-up, the
mean HbA1c levels equalized rapidly between the 2 glycemia control arm of the study was discontinued because
originally randomized groups, with that of the original of a statistically significant increase in all-cause mortality in
intense treatment group rising to approximately 8% and the intensive therapy group compared with the standard
that of the standard group falling to the same level. group (5.0% and 4.0%, respectively; hazard ratio, 1.22;
Despite this equalization of glycemic control in the 2 95% confidence interval, 1.01e1.46; P ¼ 0.04).7 The risk
groups, there continued to be an approximately 50% risk for CVD mortality was increased by 35% (P ¼ 0.02).
reduction of further retinopathy progression in the original Although there was an increased risk of mortality in the
intensive control group, a phenomenon termed metabolic intensive therapy group, there was a statistically
memory.4 Even 3 decades after the original randomization, significant approximately one-third reduction in risk of
the relative benefits from being assigned to the intensive progression of retinopathy for various definitions of reti-
treatment group continue to accrue, with an almost 50% nopathy progression in the intensive therapy group
reduction in the risk of advanced retinal complications and compared with the standard therapy group at the end of the
of having ocular surgery (primarily cataract, vitrectomy, ACCORD study.8 As with the EDIC study, although the
and retinal detachment surgery) in the original intensive HbA1c levels quickly became equivalent in the 2
treatment group.5 Lest one think this is important only in randomly assigned groups after that arm of the study
persons with type 1 diabetes, the 10-year posttrial follow- was discontinued, those participants originally assigned
up of participants with type 2 diabetes in the United to the intensive therapy group continued to have reduced
Kingdom Prospective Diabetes Project (UKPDS), who rates of retinopathy. In the 4 years after the conclusion
initially were randomly assigned to intensive glucose ther- of the ACCORD trial, retinopathy progressed by 3 steps
apy, also showed a reduction in risk for microvascular or more in the Early Treatment Diabetic Retinopathy
complications compared with those initially assigned to scale in 5.8% of the original intensive treatment group
conventional care.6 The continued effect of the original versus 12.7% in the standard treatment group (adjusted

1840 Ó 2016 Published by Elsevier on behalf of the American Academy of Ophthalmology http://dx.doi.org/10.1016/j.ophtha.2016.05.039
ISSN 0161-6420/16
 Editorial

odds ratio, 0.42; 95% confidence interval, 0.28e0.63; patients with CVD seems to be a sensible goal. The options
P < 0.0001).1 to achieve this metabolic goal have continued to expand. In
How can we reconcile the apparently adverse effects of type 1 diabetes, the use of continuous glucose monitoring
intensive therapy on cardiovascular mortality in the may prove to be useful, especially as a safeguard against
ACCORD study and the repeatedly demonstrated benefits of hypoglycemia in those with hypoglycemia unawareness and
intensive therapy on retinopathy and other microvascular as an essential component in the development of the artifi-
complications? What should we tell our patients? cial pancreas.12 In type 2 diabetes, a panoply of new
First, it is clear in type 1 and type 2 diabetes that inten- medications have been introduced since the end of the
sive therapy achieving a mean HbA1c of approximately 7% UKPDS. Although their relative roles are still being
results in substantial benefits in preventing and delaying the defined,13 the expanded menu of glucose-lowering drugs
progression of retinopathy, nephropathy, and neuropathy. provides many new options for achieving the glycemic
The ACCORD and ACCORDION findings support an goals that have been shown to reduce the long-term com-
additional benefit with HbA1c levels that are even lower plications of diabetes mellitus.
than 7%, as has been suggested by the DCCT.9 However, The results of numerous well-designed clinical trials and
whether the costs and risks, especially with regard to long-term follow-up studies have demonstrated the benefits
hypoglycemia, of achieving HbA1c levels are much lower of controlling glycemia. The increasing options for patients
than 7% with current day therapies are balanced by the both to monitor and control their blood glucose levels
additional benefits is unclear. Therefore, the current goal should improve our ability to reach the currently recom-
of aiming for an HbA1c level of less than 7%, but not mended HbA1c level goal of less than 7% safely and cost
less than 6%, for most patients seems justified. The effectively. Although intravitreal anti-VEGF treatment has
glycemic goals need to be individualized based on been demonstrated to slow the progression of, and in
expected longevity, comorbidities, and other factors. some cases even reverse, retinopathy severity in eyes with
Second, the role of intensive therapy on the less diabetes- diabetic macular edema or proliferative diabetic retinopathy,
specific CVD complications is less clear. Both the DCCT it is apparent that intervening early to prevent or slow
and UKPDS, which remain the iconic, driving studies sup- retinopathy development by controlling glycemia, blood
porting intensive therapy, demonstrated beneficial effects of pressure, and serum lipidemia is the first-line public
intensive therapy on CVD outcomes.6,10 The specific med- health approach.14,15
ications used in the UKPDS and DCCT, both of which
included patients with little if any CVD at baseline, included
insulin, metformin and sulfonylureas and insulin, respec- References
tively. The main consideration should be what was different
in the ACCORD study, especially compared with the
UKPDS, which may have resulted in an adverse effect of 1. Action to Control Cardiovascular Risk in Diabetes Follow-On
intensive therapy on CVD. In truth, we do not know; (ACCORDION) Eye Study Group and the Action to Control
Cardiovascular Risk in Diabetes Follow-On (ACCORDION)
however, notably, the ACCORD study included patients
Study Group. Persistent effects of intensive glycemic control
with a longer duration of type 2 diabetes than those in the on retinopathy in type 2 diabetes in the Action to Control
UKPDS and with a far higher prevalence of CVD at base- Cardiovascular Risk in Diabetes (ACCORD) follow-on study.
line than in either DCCT or UKPDS.6,10 The ACCORD Diabetes Care 2016;39:1089–100.
study included only 3 years of follow-up compared with 2. Diabetes Control and Complications Trial Research Group.
approximately 20 years for DCCT and UKPDS. The effect of intensive therapy of diabetes on the develop-
The ACCORD study therefore was more of a short-term ment and progression of long-term complications in insulin-
secondary intervention study, and its results may have dependent diabetes mellitus. N Engl J Med 1993;329:
little to do with the results when intensive therapy is applied 977–86.
earlier in the course of diabetes, which is recommended 3. Nathan DM, Bayless M, Cleary P, et al. The Diabetes Control
and Complications Trial/Epidemiology of Diabetes
widely based on its salutary effects on microvascular dis-
Interventions and Complications Study at 30 years: advances
ease. Secondary analyses have suggested that the intensively and contributions. Diabetes 2013;62:3976–86.
treated patients who were adversely affected had high 4. Diabetes Control and Complications Trial/Epidemiology of
HbA1c levels, that is, they were exposed to the potential Diabetes Interventions and Complications Research Group.
adverse effects of the intensive regimen without the benefits Prolonged effect of intensive therapy on the risk of retinopathy
that accrue with lower HbA1c levels.11 Finally, intensive complications in patients with type 1 diabetes mellitus, 10
therapy in the ACCORD study required as many as 5 years after the Diabetes Control and Complications Trial. Arch
glucose-lowering medications, with most patients using at Ophthalmol 2008;126:1707–15.
least 3. This multidrug treatment is in stark contrast with the 5. Aiello LP, Sun W, Das A, et al. Intensive diabetes therapy and
number of drugs used in the UKPDS and in practice, which ocular surgery in type 1 diabetes. N Engl J Med 2015;372:
1722–33.
is usually no more than 2.
6. Holman R, Paul S, Bethel A, et al. 10-year follow-up of
Based on the balance of demonstrated benefits and risks intensive glucose control in type 2 diabetes. N Engl J Med
shown in the EDIC study, UKPDS, and the ACCORD study 2008;359:1577–89.
and taking into account the increased risk for mortality 7. The Action to Control Cardiovascular Risk in Diabetes Study
shown in the ACCORD study, aiming for an HbA1c level of Group. Effects of intensive glucose lowering in type 2 dia-
less than 7% and avoiding diabetes polypharmacy in betes. N Engl J Med 2008;358:2445–559.

1841
 Ophthalmology Volume 123, Number 9, September 2016

8. Chew EY, Davis MD, Danis RP, et al. The effects of medical mortality during a median 3.4-year follow-up of glyce-
management on the progression of diabetic retinopathy in mic treatment in the ACCORD trial. Diabetes Care
persons with type 2 diabetes: the Action to Control Cardio- 2010;33:983–90.
vascular Risk in Diabetes (ACCORD) eye study. Ophthal- 12. Russell SJ, El-Khatib FH, Sinha M, et al. Outpatient glycemic
mology 2014;121:2443–51. control with a bionic pancreas in type 1 diabetes. N Engl J
9. Lachin JM, Genuth S, Nathan DM, et al. The effect of gly- Med 2014;371:313–25.
cemic exposure on the risk of microvascular complications in 13. Nathan DM, Buse JB, Kahn SE, et al. Rationale and design of
the Diabetes Control and Complications Trialdrevisited. the glycemia reduction approaches in diabetes: a comparative
Diabetes 2008;57:995–1001. effectiveness study (GRADE). Diabetes Care 2013;36:2254–61.
10. The Diabetes Control and Complications Trial/Epidemiology 14. Ip M, Domalpally A, Hopkins J, et al. Long-term effects of
of Diabetes Interventions and Complications (DCCT/EDIC) ranibizumab on diabetic retinopathy severity and progression.
Study Research Group. Intensive diabetes treatment and car- Arch Ophthalmol 2012;130:1145–52.
diovascular disease in patients with type 1 diabetes. N Engl J 15. Diabetic Retinopathy Clinical Research Network. Panretinal
Med 2005;353:2643–53. photocoagulation vs intravitreous ranibizumab for proliferative
11. Riddle MC, Ambrosius WT, Brillon DJ, et al. Epide- diabetic retinopathy: a randomized trial. JAMA 2015;314:
miologic relationships between A1C and all-cause 2137–46.

Footnotes and Financial Disclosures

1842