Glycated hemoglobin variability andretinopathy progression in type 1 diabetes:Is month-to-moth instability a better predictor?

Glucose instability has been proposed tobe one of the risk factors for diabeticcomplications, in particular retinopathy,since the publication of the DiabetesControl and Complication Trial (DCCT)20 years ago. In that study, as comparedwith conventional therapy, intensive ther-apy effectively delays the onset and slowsthe progression of diabetic complications

in patients with type 1 diabetes1. Thefollowing analyses have found that inaddition to mean glycated hemoglobin(HbA1c) level, high HbA1c variability, butnot glucose variability, might be associ-ated with increased risks of retinopathy,nephropathy and perhaps neuropathy, intype 1 diabetic patients (Table 1)2–8.Although many laboratory and clinical

studies suggested that ‘short-term’ glucosefluctuation, mostly driven by high post-prandial blood glucose levels, leading toincreased oxidative stress and decreasedendothelial function, is responsible forthe higher risks of microvascular andcardiovascular outcomes in type 2 dia-betic patients9, the role of ‘longer-term’glucose stability in terms of HbA1c

Table 1 | Studies examining the relationship between glucose or glycated hemoglobin variability and long-term complications in type 1 diabeticpatients

Author, year Data source No.participants

Follow-upduration

Variability measurement Outcome Results

Kilpatrick, 20062 DCCT 1,441 5 years Seven-point glucoseprofiles were collectedquarterly. The meanarea under the curveand the SD of glucosevariability within-dayand between visitswere calculated

RetinopathyNephropathy

Within-day and between-dayglucose variability does notappear to be an additionalfactor in the developmentor progression of eitherretinopathy nor nephropathy

Kilpatrick, 20083 DCCT 1,441 NR HbA1c was collectedquarterly. The meanHbA1c and the SD ofHbA1c variability from6 months onwardswere calculated

RetinopathyNephropathy

A1c variability added to meanA1c in predicting the risk ofdevelopment or progressionof retinopathy (HR 2.26 forevery 1% increase in HbA1cSD [95% CI 1.63–3.14]) andnephropathy (HR 1.80 [95% CI1.37–2.42])

Bragd, 20084 Outpatientclinic at onehospital

100 11 years SD of blood glucosewas calculated basedon 70 measurementsover 4 weeks

RetinopathyNephropathyNeuropathy

Glucose variability wassignificantly related to thepresence of peripheralneuropathy (OR 2.34 [96% CI1.06–5.20]), but not withincidence of proliferativeretinopathy (HR 1.37 [95% CI0.77–2.43]) nor nephropathy(HR 1.75 [95% CI 0.89–3.40])

*Corresponding author. Lee-Ming ChuangTel.: +886-2-23123456 ext. 65038 Fax: +886-2-23938859E-mail address: leeming@ntu.edu.twReceived 21 July 2013; revised 7 August 2013;accepted 9 August 2013

ª 2013 The Authors. Journal of Diabetes Investigation published by Asian Association for the Study of Diabetes (AASD) and Wiley Publishing Asia Pty Ltd J Diabetes Invest Vol. 5 No. 2 March 2014 149This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distributionin any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

COMMENTARY

variability, and the occurrence and pro-gression of diabetic complications amongtype 1 diabetic patients remains elusive.In a recently published study, Hietala

et al.10 analyzed data from 2,019 type 1diabetic patients who were enrolled in theFinnish Diabetic Nephropathy (Finn-Diane) study. During the average followup of 5.2 years, they have reconfirmedthat high HbA1c variability was associatedwith an increased risk of incidence, as wellas laser treatment for retinopathy. In theearlier report, they also showed thatHbA1c variability was not only predictiveof incident microalbuminuria and pro-

gression of renal disease, but also of inci-dent cardiovascular events in type 1diabetic patients7. Of note, patients withhigh mean HbA1c values also had highHbA1c variability. However, patients withlow mean HbA1c value, but high HbA1c

variability, had similar incidence fornephropathy progression as patients withhigh mean HbA1c value, but low HbA1c

variability, suggesting an equally impor-tant effect of both absolute value and sta-bility in the risk of progression in renaloutcome. The significant associationbetween HbA1c variability, and risks ofretinopathy and nephropathy persisted

even after taking the number of HbA1c

measurements into consideration.Why was HbA1c instability indepen-

dently associated with a higher chance oflaser treatment for retinopathy? Theauthors emphasized that even short peri-ods of higher HbA1c could have a sus-tained effect on worsening of retinopathydespite subsequent improvement in gly-cemic control10. However, because ofthe inherent limitations in the observa-tional study design, there might be otherexplanations for the observed association(Figure 1). First, because there might bea lag time between worsening of visual

Table 1 (Continued)

Author, year Data source No.participants

Follow-upduration

Variability measurement Outcome Results

Kilpatrick, 20095 DCCTEDIC (first 4years)

1,441 NR Mean area under thecurve glucose andthe within-day glucosevariability (SD andMAGE) during theDCCT

RetinopathyNephropathy

Glucose variability in the DCCTdid not predict thedevelopment of retinopathyor nephropathy byEDIC year 4

Siegelaar, 20096 DCCT 1,160 5 years Variability of bloodglucose (within-day SD)was calculated as theSD of daily bloodglocose around themean from eachquarterly visit andMAGE

Neuropathy No effect of glucose variabilityon confirmed clinicalneuropathy, autonomicneuropathy and abnormalnerve conduction

Waden, 20097 FinnDiane 2,107 5.7 years SD of all recorded A1c Incidentmicroalbuminuria,renal diseaseprogression,cardiovascular event(MI, stroke, CABG,PTCA, limb ampuation)

SD of serial A1c wasindependently associatedwith progression of renaldisease (HR 1.92 [95% CI1.49–2.47]) and of acardiovascular event (HR 1.98[95% CI 1.39–2.82])

Marcovecchio,20118

ORPSNFS

1,232 NR For each patient, theintrapersonal SD forannual HbA1c wascalculated

Microalbuminuriadevelopment

HbA1c SD was independentlyassociated withmicroalbuminuria (HR 1.31 forevery 1% increase in HbA1cSD [95% CI 1.01-1.35])

Hietala, 201310 FinnDiane 2,019 5.2 years Coeffecient of variation(the ratio of intrapersonalSD and mean) of HbA1c

Retinopathy HbA1c variability was associatedwith an increased risk ofretinopathy requiring lasertreatment (HR 1.60 [95% CI1.10–2.50] comparing thehighest with the lowestquartile)

CABG, coronary artery bypass graft surgery; CI, confidence interval; DCCT, Diabetes Control and Complication Trial; EDIC, Epidemiology of DiabetesInterventions and Complications; HbA1c, glycated hemoglobin; HR, hazard ratio; MAGE, mean amplitude of glycemic excursions; MI, myocardialinfarction; NR, not reported; OR, odds ratio; PTCA, percutaneous transluminal coronary angioplasty; SD, standard deviation.

150 J Diabetes Invest Vol. 5 No. 2 March 2014 ª 2013 The Authors. Journal of Diabetes Investigation published by AASD and Wiley Publishing Asia Pty Ltd

C OMMEN T A R Y

HbA1c variability and diabetic retinopathy http://onlinelibrary.wiley.com/journal/jdi

symptoms and laser treatment, it is pos-sible that diagnosis of retinopathy pro-gression per se leads to attempts toimprove glycemic control by physicians,and hence increasing HbA1c fluctuation(protopathic bias or reverse causation).Second, for poorly controlled patientswith high and unstable HbA1c value,physicians were more likely to do fundo-scopic or fundus photographic examina-tions, and were hence more likely to findretinopathy progression that mightrequire laser treatment (detection bias).Third, from prior studies we knew that avariable HbA1c profile is associated withseveral factors, such as low socioeco-nomic status, physical inactivity andsmoking7. Perhaps, it is these lifestyleand behavior factors (including medica-tion non-adherence) that lead to pooroutcomes.Again, we want to reassure our patients

that long-term optimal glycemic control iscrucial to reduce retinopathy occurrenceand its complications. For healthcare pro-fessionals who are involved in the care ofdiabetic patients, the findings of the studyby Hietala et al.10 highlight that patientswith highly variable HbA1c warrant carefulmonitoring of retinopathy developmentand progression. Further research on the

mechanism of the observed associationand how to prevent early retinopathyworsening for those poorly controlledtype 1 patients undergoing intensive glu-cose lowering therapy are required to fillthe knowledge gap.

Chia-Hsuin Chang, Lee-Ming Chuang*Department of Internal Medicine,

National Taiwan University Hospital, andInstitute of Preventive Medicine, National

Taiwan University School of PublicHealth, Taipei, Taiwan

REFERENCES1. The Diabetes Control and

Complications Trial Research Group.The effect of intensive treatment ofdiabetes on the development andprogression of long-termcomplications in insulin-dependentdiabetes mellitus. N Eng J Med 1993;329: 977–986.

2. Kilpatrick ES, Rigby AS, Atkin SL. Theeffect of glucose variability on therisk of microvascular complications intype 1 diabetes. Diabetes Care 2006;29: 1486–1490.

3. Kilpatrick ES, Rigby AS, Atkin SL.A1c variability and the risk ofmicrovascular complications in type 1

diabetes. Diabetes Care 2008; 31:2198–2202.

4. Bragd J, Admson U, Backlund LB,et al. Can glycaemic variability, ascalculated form blood glucose self-monitoring, predict the developmentof complications in type 1 diabetesover a decade? Diabetes Metab 2008;34: 612–616.

5. Kilpatrick ES, Rigby AS, Atkin SL. Effectof glucose variability on the long-term risk of microvascularcomplications in type 1 diabetes.Diabetes Care 2009; 32: 1901–1903.

6. Siegelaar SE, Kilpatrick ES, Rigby AS,et al. Glucose variability does notcontribute to the development ofperipheral and autonomicneuropathy in type 1 diabetes: datafrom the DCCT. Diabetologia 2009; 52:2229–2232.

7. Waden J, Forsblom C, Thorn LM,et al. A1c variability predictsincident cardiovascular events,microalbuminuria, and over diabeticnephropathy in patients with type 1diabetes. Diabetes 2009; 58:2649–2655.

8. Marcovecchio ML, Dalton RN, ChiarelliF, et al. A1c variability as anindependent risk factor for

Daily glucose variability

↑ Treatment↑ Screening

AgeVisual

symptoms

Detection biasConfounding by behavior factors

Unhealthy dietPhysical inactivity

SmokingMedication non-adherence

Confounding by disease severityDiabetic duration

Insulin insensitivity

HbA1C variability1Diabetic

retinopathy HbA1C variability 2

Reversecausation

Figure 1 | Possible explanations for the observed positive association between glycated hemoglobin (HbA1c) variability and retinopathy amongtype 1 diabetic patients.

ª 2013 The Authors. Journal of Diabetes Investigation published by AASD and Wiley Publishing Asia Pty Ltd J Diabetes Invest Vol. 5 No. 2 March 2014 151

C OMM EN T A R Y

http://onlinelibrary.wiley.com/journal/jdi HbA1c variability and diabetic retinopathy

microalbuminuria in young peoplewith type 1 diabetes. Diabetes Care2011; 34: 1011–1013.

9. Nalysnyk L, Hernandez-Medina M,Krishnarajah G. Glycaemic variabilityand complications in patients with

diabetes mellitus: evidence from asystematic review of the literature.Diabetes Obes Metab 2010; 12: 288–289.

10. Hietala K, Waden J, Forsblom C, et al.HbA1c variability is associated with an

increased risk of retinopathy requiringlaser treatment in type 1 diabetes.Diabetologia 2013; 56: 737–745.

Doi: 10.1111/jdi.12147

152 J Diabetes Invest Vol. 5 No. 2 March 2014 ª 2013 The Authors. Journal of Diabetes Investigation published by AASD and Wiley Publishing Asia Pty Ltd

C OMMEN T A R Y

HbA1c variability and diabetic retinopathy http://onlinelibrary.wiley.com/journal/jdi