DIABETICMedicine DOI: 10.1111/j.1464-5491.2011.03372.x
Article: Complications Corneal confocal microscopy detects improvement in corneal nerve morphology with an improvement in risk factors for diabetic neuropathy M. Tavakoli, P. Kallinikos, A. Iqbal, A. Herbert*, H. Fadavi, N. Efron†, A. J. M. Boulton and R. A Malik Division of Cardiovascular Medicine, University of Manchester and Manchester Royal Infirmary, *Pennine Acute Hospitals NHS Trust, Manchester, UK and †Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Queensland, Australia Accepted 20 June 2011
Abstract We have assessed whether corneal confocal microscopy can be used to detect alterations in nerve morphology following an improvement in risk factors associated with diabetic neuropathy.
Aim
Methods Twenty-five patients with diabetes with mild to moderate neuropathy and 18 control subjects underwent corneal confocal microscopy to quantify corneal nerve fibre (density, branch density, length and tortuosity) at baseline and after 24 months from first visit. This was not planned as an intervention trial and was simply an observational follow-up.
At baseline, nerve fibre density (18.8 2.1 vs. 46.0 3.8 number ⁄ mm2, P = 0.001), nerve branch density (6.9 1.5 vs. 35.6 6.7 number ⁄ mm2, P < 0.0001), nerve fibre length (8.3 0.9 vs. 13.5 0.8 mm ⁄ mm2, P < 0.0001) and nerve fibre tortuosity (19.8 1.6 vs. 22.7 2.2, P < 0.05) were significantly lower in patients with diabetes than in control subjects. At follow-up, glycaemic control (HbA1c 64 3 to 58 2 mmol ⁄ mol, P = 0.08), total cholesterol (4.9 0.2 to 4.2 0.2 mmol ⁄ l, P = 0.01), systolic blood pressure (145.8 4.9 to 135.9 3.7 mmHg, P = 0.09) and diastolic blood pressure (77.8 2.7 to 70.8 2.5, P = 0.03) improved. Nerve fibre density (24.1 2.0, P = 0.05), nerve branch density (11.1 1.3, P < 0.01) and nerve fibre tortuosity (22.6 1.5, P = 0.05) increased significantly, with no change in nerve fibre length (8.4 0.5). Improvement in nerve fibre density correlated significantly with the improvement in HbA1c (r = )0.51, P = 0.008). Via four multifactorial regressions, this confirms the negative association between HbA1c and nerve fibre density (P = 0.02). Results
Conclusions This study shows that corneal confocal microscopy may be employed in longitudinal studies to assess progression of human diabetic neuropathy and also supports the hypothesis that improvements in risk factors for diabetic neuropathy, in particular HbA1c, may lead to morphological repair of nerve fibres.
Diabet. Med. 28, 1261–1267 (2011) Keywords corneal confocal microscopy, corneal nerves, diabetic neuropathy, risk factors
Introduction Diabetic polyneuropathy is among the most common long-term complications of diabetes and is the main precipitating factor for foot ulceration and lower extremity amputation [1]. Furthermore, a clinical history of diabetic neuropathy has recently been shown to be one of only three independent risk factors to predict mortality in patients with Type 2 diabetes [2].
Correspondence to: Professor R. A. Malik, Division of Cardiovascular Medicine, University of Manchester, Manchester M13 9NT, UK. E-mail:
[email protected]
ª 2011 The Authors. Diabetic Medicine ª 2011 Diabetes UK
Longitudinal studies are limited, but show a deterioration of neuropathy over time in both Type 1 [3] and Type 2 diabetes [4], and the placebo arms of several clinical trials in diabetic neuropathy show a monotonic worsening in electrophysiology and quantitative sensory testing [5]. The Diabetes Control and Complications Trial and the recent Epidemiology of Diabetes Interventions and Complications study have confirmed the immediate and durable benefit of improved glycaemic control on neuropathy in patients with Type 1 diabetes [3]. However, in Type 2 diabetes, the Veterans Affairs Diabetes Trial demonstrated no impact of improved glycaemic control on somatic neuropathy and, in fact, a slight
1261
DIABETICMedicine
Corneal confocal microscopy detects improvement in diabetic neuropathy • M. Tavakoli et al.
worsening of autonomic neuropathy [6]. Furthermore, multiple clinical trials in diabetic neuropathy have demonstrated limited or no efficacy of various pathogenetically relevant interventions [1]. These data have prompted several investigators to question the reliability of the endpoints currently employed in trials of human diabetic neuropathy [5]. This has led to the suggestion that the lack of significant improvement in clinical trials of diabetic neuropathy is because of the lack of deterioration in the placebo arm of several clinical trials, possibly attributable to an improvement in the risk factors for human diabetic neuropathy [5]. The link between neuropathy and cardiovascular risk factors, such as lipids, blood pressure and BMI, has been established in a study of patients with Type 1 diabetes followed over 7 years [7]. Whilst multiple risk factor intervention has demonstrated a significant improvement in retinopathy, nephropathy and autonomic neuropathy, no benefit was seen for somatic neuropathy [8]. Similarly, in a lifestyle intervention study to improve weight, lipids, blood pressure and glycaemia in subjects with impaired glucose tolerance, vibration perception and electrophysiology did not improve, whilst quantitative sudomotor axon reflex test and intra-epidermal nerve fibre density, a functional and structural measure of small fibre damage, did [9]. Thus, quantification of small fibre pathology may be a key to assessing degeneration and regeneration of nerves in patients with diabetes. However, techniques such as nerve and skin biopsy are invasive and cannot be readily deployed in natural history or intervention studies. Recently, we have shown that corneal confocal microscopy accurately detects corneal small nerve fibre damage, which is directly related to the level of severity of neuropathy [10–12] and intra-epidermal nerve fibre density in skin biopsy [13]. Furthermore, corneal confocal microscopy can detect nerve
(a)
(b)
FIGURE 1 Corneal confocal microscopy images of Bowman’s layer from a diabetic patient at (a) baseline and (b) follow-up.
Table 1 Clinical detail of control subjects and patients with diabetes at baseline and at 24 months
n Age (years) Sex (female ⁄ male) Duration of diabetes Type of diabetes (Type 1 ⁄ Type 2) Neuropathy Impairment Score Vibration perception threshold (volts) HbA1c (mmol ⁄ mol) HbA1c (%) Total cholesterol (mmol ⁄ l)à Triglycerides (mmol ⁄ l) HDL (mmol ⁄ l) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Nerve fibre density (number ⁄ mm2)* Nerve branch density (number ⁄ mm2)à Nerve fibre length (mm ⁄ mm2) Nerve fibre tortuosity (tortuosity coefficient)*
Control
Baseline
18 57.0 8 ⁄ 10 — — 0 9.3 45.6 5.7 — — — — — 46.0 35.6 13.5 22.7
25 52.0 5 ⁄ 20 26.5 15 ⁄ 10 10.3 16.86 64.60 8.1 4.9 1.8 1.4 145.8 77.8 18.8 6.9 8.3 19.8
3.0
2.6 1.11 0.1
3.8 6.7 0.8 2.2
Follow-up (24 months)
2.0 2.5 2.8 3.05 3.34 0.3 0.2 0.2 0.1 4.9 2.7 2.1 1.5 0.9 1.6
25 54.0 5 ⁄ 20 28.5 15 ⁄ 10 — — 58.72 7.5 4.2 1.5 1.3 135.9 70.8 24.1 11.1 8.4 22.6
2.0 2.5
2.53 0.2 0.2 0.2 0.1 3.7 2.5 2.0 1.3 0.5 1.5
Significant improvement at follow-up compared with baseline: *P £ 0.05; P £ 0.03; àP £ 0.01.
1262
ª 2011 The Authors. Diabetic Medicine ª 2011 Diabetes UK
Original article
fibre repair within 6 months of pancreas transplantation [14]. To test the hypothesis that corneal confocal microscopy may be useful in monitoring the progression of human diabetic neuropathy, we have evaluated corneal nerve fibre morphology at baseline and follow-up after approximately 24 months in relation to change in glycaemia, lipids and blood pressure.
Subjects and methods Twenty-five patients with diabetes attending the Manchester Diabetes Centre and 18 healthy volunteers without diabetes underwent assessment of neuropathy using the Neuropathy Impairment Score and vibration perception threshold. Patients were excluded if they had a neuropathy of any other cause, absent pedal pulses, wore contact lenses or had a history of corneal trauma or surgery. The protocol was approved by the local research ethics committee of the Greater Manchester Health Authority and all subjects gave written informed consent. This study was not planned as an intervention trial and was simply an observational follow-up. The patients underwent corneal confocal microscopy examination with a Tomey Confoscan confocal microscope Model P4 (Erlangen, Germany) applying our established methodology [12,13]. Several scans of the entire depth of the cornea were recorded and 3–5 high-quality images with best
DIABETICMedicine
resolution of the sub-basal nerve plexus were acquired from the centre of the cornea. The investigator who examined the cornea and who undertook morphometric measurements of the images was masked with respect to the identity of the patients and medical and neurological results of subjects. The following variables were quantified to define corneal nerve fibre damage and repair: (1) nerve fibre density—the total number of major nerves per mm2; (2) nerve fibre length—the total length of all nerve fibres and branches per mm2; (3) nerve branch density—the number of branches emanating from each nerve trunk per mm2; (4) nerve fibre tortuosity is mathematically derived from the images [10]. Measures 1 and 3 were determined using morphometric software incorporated within the Tomey instrument, measure 2 was determined using third-party image analysis software (Scion Image for Windows; Scion Corporation, Frederick, MA., USA) and measure 4 was calculated using a MATLAB function (Mathworks, version 6.5; MATLAB, Mathworks, USA) that was created for this purpose [10]. To estimate the error in measuring nerve fibre density, nerve fibre length and nerve branch density, we acquired images and determined each of these variables on 15 subjects on two occasions, separated by at least 48 h. The coefficient of variation of these measures was: 12% for nerve fibre density, 9% for nerve fibre length and 24% for nerve branch density.
FIGURE 2 Change in HbA1c (P = 0.08), cholesterol (P = 0.01), systolic (P = 0.09) and diastolic (P = 0.03) blood pressure from baseline to follow-up.
ª 2011 The Authors. Diabetic Medicine ª 2011 Diabetes UK
1263
Corneal confocal microscopy detects improvement in diabetic neuropathy • M. Tavakoli et al.
DIABETICMedicine
Statistical analysis
Baseline
SPSS 11.05.0 for Windows (SPSS Inc., Chicago, IL, USA) was used to compute the results. Data are expressed as mean sem and the analysis includes descriptive and frequency statistics. ANOVA with Scheffe post-hoc tests was used to study differences between means. The association between the change in various risk factors and nerve fibre density, nerve fibre length, nerve branch density and nerve fibre tortuosity as the dependent variables were assessed by use of a multifactorial regression, via the regression coefficients and their confidence intervals.
At baseline, nerve fibre density (18.8 2.1 vs. 46.0 3.8 number ⁄ mm2, P = 0.001), nerve branch density (6.9 1.5 vs. 35.6 6.7 number ⁄ mm2, P < 0.0001), nerve fibre length (8.3 0.9 vs. 13.5 0.8 mm ⁄ mm2, P < 0.0001) and nerve fibre tortuosity (19.8 1.6 vs. 22.7 2.2, P < 0.05) were significantly reduced in subjects with diabetes compared with control subjects (Table 1).
Follow-up Clinical metabolic variables
Results Twenty-five people with diabetes (15 with Type 1 and 10 with Type 2) aged 52 2 years with mild to moderate neuropathy (Neuropathy Impairment Score 10.3 2.8 and vibration perception threshold 16.86 3.05) and 18 age-matched (57.0 3.0 years) control subjects were studied. The people with diabetes were attending a specialist diabetes centre to improve overall glycaemic control and cardiovascular risk factors. Corneal nerve morphology and glycaemic control (HbA1c), total cholesterol, HDL, triglycerides, systolic and diastolic blood pressure were assessed at baseline and after 24 months. This was not planned as an intervention trial and was simply an observational follow-up.
(a)
(c)
An improvement was demonstrated from baseline to follow-up in glycaemic control (HbA1c 64.60 3.34 to 58.72 2.53 mmol ⁄ mol, P = 0.08), total cholesterol (4.9 0.2 to 4.2 0.2 mmol ⁄ l, P = 0.01), systolic blood pressure (145.8 4.9 to 135.9 3.7 mmHg, P = 0.09) and diastolic blood pressure (77.8 2.7 to 70.8 2.5 mmHg, P = 0.03), with a non significant decrease in triglycerides (1.8 0.2 vs. 1.5 0.2 mmol ⁄ l) and no change in HDL (1.4 0.1 vs. 1.3 0.1 mmol ⁄ l) (Table 1, Fig. 1). Corneal nerve morphology
Nerve fibre density (18.8 2.1 to 24.1 2.0, P < 0.05), nerve branch density (6.9 1.5 to 11.1 1.3, P < 0.01) and nerve fibre tortuosity (19.8 1.6 to 22.6 1.5, P < 0.05) increased
(b)
(d)
FIGURE 3 Change in corneal confocal microscopy corneal nerve morphology from baseline to follow-up.
1264
ª 2011 The Authors. Diabetic Medicine ª 2011 Diabetes UK
Original article
DIABETICMedicine
[Pearson’s correlation coefficient = )0.51 (95% CI )0.76 to )0.13), P = 0.008] (Fig. 4). To assess the role of a change in all the risk factors under consideration, a multifactorial regression was carried out in turn for each of the corneal confocal microscopy variables, whereby the outcome of this regression was defined as the change from baseline to 24 months in each, and the potential risk factors for neuropathy progression, which were included in the regression simultaneously. The change from baseline to 24 months of the variables and the type of diabetes was inserted into the regression analyses to adjust estimates for potential confounding factors, such as type of diabetes. The results of these four regressions are shown in Table 2 and demonstrate a significant association between HbA1c and nerve fibre density (coefficient )3.4, 95% CI )6.24 to )0.57, P-value = 0.02).
Discussion FIGURE 4 Correlation between change in nerve fibre density with change in HbA1c (r = )0.52, P = 0.008).
significantly, with no change in nerve fibre length (8.3 0.9 to 8.4 0.5) (Table 1, Figs 2 and 3). Correlation ⁄ regression between risk factors and corneal confocal microscopy
The increase in nerve fibre density was significantly associated with the reduction in HbA1c from baseline to follow-up
The natural history of nerve damage in people with diabetes is not determined. Longitudinal data from the Rochester cohort supports the contention that the duration and severity of hyperglycaemia are related to the severity of neuropathy [15]. In a study of people with Type 2 diabetes, 21% developed a significant neuropathy over 4 years [16]. In a long-term followup study of another cohort of people with Type 2 diabetes, nerve conduction abnormalities in the legs and feet increased from 8% at baseline to 16% after 5 years and to 42% after 10 years [4]. However, it is important to note that each of these studies used
Table 2 Regression estimates for risk factors, with outcome as change in nerve fibre density, nerve branch density, nerve fibre length and nerve fibre tortuosity between baseline and 24 months Risk factor Nerve fibre density HbA1c (%) Total cholesterol (mmol) Triglyceride (mmol) HDL (mmol) Systolic blood pressure (mmHg) Nerve branch length HbA1c (%) Total cholesterol (mmol) Triglyceride (mmol) HDL (mmol) Systolic blood pressure (mmHg) Nerve fibre length HbA1c (%) Total cholesterol (mmol) Triglyceride (mmol) HDL (mmol) Systolic blood pressure (mmHg) Nerve fibre tortuosity HbA1c (%) Total cholesterol (mmol) Triglyceride (mmol) HDL (mmol) Systolic blood pressure (mmHg)
Coefficient*
95% CI
P-value
)3.40 )0.64 )1.28 2.68 )0.10
)6.24 )5.41 )6.20 )11.81 )1.06
to to to to to
)0.57 4.12 3.65 17.16 0.31
0.02 0.78 0.59 0.70 0.31
)0.81 0.98 0.74 )2.84 )0.01
)4.39 )5.04 )5.48 )21.13 )0.27
to to to to to
2.77 6.99 6.95 15.44 0.25
0.64 0.74 0.80 0.75 0.95
)329.74 )619.79 2.91 2389.66 16.68
)1478.02 )2548.23 )1990.29 )3475.09 )66.65
to to to to to
818.54 1308.65 1996.11 8254.41 100.01
0.55 0.51 0.998 0.40 0.68
0.28 0.14 )1.28 )2.90 0.07
)1.68 )3.16 )4.69 )12.93 )0.08
to to to to to
2.25 3.43 2.13 7.14 0.21
0.77 0.93 0.44 0.55 0.35
*Estimates adjusted for effect of type of diabetes.
ª 2011 The Authors. Diabetic Medicine ª 2011 Diabetes UK
1265
DIABETICMedicine
Corneal confocal microscopy detects improvement in diabetic neuropathy • M. Tavakoli et al.
different clinical tests to evaluate neuropathy and the focus was on large fibres. A number of metabolic risk factors, such as glucose, lipids, blood pressure and BMI, have been shown to be related to the development of diabetic neuropathy [7,17,18]. An improvement in glycaemic control improves diabetic neuropathy in people with Type 1 diabetes [3] but may have no [6] or minimal benefit [19] in people with Type 2 diabetes. Whilst combined improvement in weight, glycaemic control, lipids and blood pressure has shown significant improvements in retinopathy, nephropathy and autonomic neuropathy, no improvement was shown in vibration perception [8,20]. Furthermore, recently lifestyle intervention, which improved weight, lipids, blood pressure and glycaemia in subjects with impaired glucose tolerance did not improve vibration perception and electrophysiology, measures of large fibre dysfunction, but did improve quantitative sudomotor axon reflex test and intraepidermal nerve fibre density, measures of small fibre dysfunction and damage, respectively [9]. Our recent studies using corneal confocal microscopy demonstrate that corneal small nerve fibre damage can be detected prior to abnormalities in electrophysiology and quantitative sensory testing [13]. Furthermore, it shows progression with the severity of neuropathy and has good sensitivity and specificity for defining those at risk of neuropathy and foot ulceration, using a relatively crude measure of neuropathic severity, the Neuropathy Deficit Score [12]. Our studies also show that corneal confocal microscopy quantifies small nerve fibre damage as accurately as intra-epidermal nerve fibre assessment using skin biopsy [13]. Added to this, we have recently shown that corneal confocal microscopy can detect early nerve fibre repair following pancreas transplantation [14]. In the present study, an improvement in HbA1c was associated with an improvement in nerve fibre density but not nerve branch density or nerve fibre tortuosity. Because of the multiple testing undertaken and the non-significance of the other risk factors with the other three corneal confocal microscopy outcomes, this requires cautious interpretation. We have no obvious explanation for the lack of this association with nerve branch density and nerve fibre tortuosity, but it is consistent with our previous data following pancreas transplantation, which also showed no improvement in these variables [14]. A major limitation of the current study is the small size of the study and also the lack of randomization and placebo control. Therefore, a larger randomized study with active intervention is required to confirm our findings. Nevertheless, the present data suggest that corneal confocal microscopy may be a convenient non-invasive technique to assess progression of nerve damage and potentially assess the effects of therapeutic intervention in future clinical trials of human diabetic neuropathy.
Competing interests
Acknowledgments
This work was supported by the Juvenile Diabetes Research Foundation International Grant 5-2002-185 and National Eye Institute Grant 1 R01 NS46259-01.
References 1 Boulton AJ, Malik RA, Arezzo JC, Sosenko JM. Diabetic somatic neuropathies. Diabetes Care 2004; 27: 1458–1486. 2 Calles-Escandon J, Lovato LC, Simons-Morton DG, Kendall DM, Pop-Busui R, Cohen RM et al. Effect of intensive compared with standard glycemia treatment strategies on mortality by baseline subgroup characteristics: the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial. Diabetes Care 2010; 33: 721– 727. 3 Albers JW, Herman WH, Pop-Busui R, Feldman EL, Martin CL, Cleary PA et al. Effect of prior intensive insulin treatment during the Diabetes Control and Complications Trial (DCCT) on peripheral neuropathy in type 1 diabetes during the Epidemiology of Diabetes Interventions and Complications (EDIC) Study. Diabetes Care 2010; 33: 1090–1096. 4 Partanen J, Niskanen L, Lehtinen J, Mervaala E, Siitonen O, Uusitupa M. Natural history of peripheral neuropathy in patients with non-insulin-dependent diabetes mellitus. N Engl J Med 1995; 333: 89–94. 5 Dyck PJ, Norell JE, Tritschler H, Schuette K, Samigullin R, Ziegler D et al. Challenges in design of multicenter trials: end points assessed longitudinally for change and monotonicity. Diabetes Care 2007; 30: 2619–2625. 6 Duckworth W, Abraira C, Moritz T, Reda D, Emanuele N, Reaven PD et al. Glucose control and vascular complications in veterans with type 2 diabetes. N Engl J Med 2009; 360: 129–139. 7 Tesfaye S, Chaturvedi N, Eaton SE, Ward JD, Manes C, IonescuTirgoviste C et al. Vascular risk factors and diabetic neuropathy. N Engl J Med 2005; 352: 341–350. 8 Gaede P, Lund-Andersen H, Parving HH, Pedersen O. Effect of a multifactorial intervention on mortality in type 2 diabetes. N Engl J Med 2008; 358: 580–591. 9 Smith AG, Russell J, Feldman EL, Goldstein J, Peltier A, Smith S et al. Lifestyle intervention for pre-diabetic neuropathy. Diabetes Care 2006; 29: 1294–1299. 10 Kallinikos P, Berhanu M, O’Donnell C, Boulton AJ, Efron N, Malik RA. Corneal nerve tortuosity in diabetic patients with neuropathy. Invest Ophthalmol Vis Sci 2004; 45: 418–422. 11 Malik RA, Kallinikos P, Abbott CA, van Schie CH, Morgan P, Efron N et al. Corneal confocal microscopy: a non-invasive surrogate of nerve fibre damage and repair in diabetic patients. Diabetologia 2003; 46: 683–688. 12 Tavakoli M, Quattrini C, Abbott C, Kallinikos P, Marshall A, Finnigan J et al. Corneal confocal microscopy: a novel noninvasive test to diagnose and stratify the severity of human diabetic neuropathy. Diabetes Care 2010; 33: 1792–1797. 13 Quattrini C, Tavakoli M, Jeziorska M, Kallinikos P, Tesfaye S, Finnigan J et al. Surrogate markers of small fiber damage in human diabetic neuropathy. Diabetes 2007; 56: 2148–2154. 14 Mehra S, Tavakoli M, Kallinikos PA, Efron N, Boulton AJ, Augustine T et al. Corneal confocal microscopy detects early nerve regeneration after pancreas transplantation in patients with type 1 diabetes. Diabetes Care 2007; 30: 2608–2612. 15 Dyck PJ, Davies JL, Litchy WJ, O’Brien PC. Longitudinal assessment of diabetic polyneuropathy using a composite score in the Rochester Diabetic Neuropathy Study cohort. Neurology 1997; 49: 229–239.
Nothing to declare.
1266
ª 2011 The Authors. Diabetic Medicine ª 2011 Diabetes UK
Original article
16 van de Poll-Franse LV, Valk GD, Renders CM, Heine RJ, van Eijk JT. Longitudinal assessment of the development of diabetic polyneuropathy and associated risk factors. Diabet Med 2002; 19: 771– 776. 17 Adler AI, Boyko EJ, Ahroni JH, Stensel V, Forsberg RC, Smith DG. Risk factors for diabetic peripheral sensory neuropathy. Results of the Seattle Prospective Diabetic Foot Study. Diabetes Care 1997; 20: 1162–1167. 18 Elliott J, Tesfaye S, Chaturvedi N, Gandhi RA, Stevens LK, Emery C et al. Large-fiber dysfunction in diabetic peripheral neuropathy is
ª 2011 The Authors. Diabetic Medicine ª 2011 Diabetes UK
DIABETICMedicine
predicted by cardiovascular risk factors. Diabetes Care 2009; 32: 1896–1900. 19 Ismail-Beigi F, Craven T, Banerji MA, Basile J, Calles J, Cohen RM et al. Effect of intensive treatment of hyperglycaemia on microvascular outcomes in type 2 diabetes: an analysis of the ACCORD randomised trial. Lancet 2010; 376: 419–430. 20 Gaede P, Vedel P, Larsen N, Jensen GV, Parving HH, Pedersen O. Multifactorial intervention and cardiovascular disease in patients with type 2 diabetes. N Engl J Med 2003; 348: 383–393.
1267
