Can J Diabetes 39 (2015) 390e397
Contents lists available at ScienceDirect
Canadian Journal of Diabetes journal homepage: www.canadianjournalofdiabetes.com
Original Research
In Vivo Corneal Confocal Microscopy and Prediction of Future-Incident Neuropathy in Type 1 Diabetes: A Preliminary Longitudinal Analysis Leif E. Lovblom MSc a, y, Elise M. Halpern MSc a, y, Tong Wu BSc a, Dylan Kelly BSc a, Ausma Ahmed MD a, Genevieve Boulet MD a, Andrej Orszag MD a, Eduardo Ng MD b, Mylan Ngo RT b, Vera Bril MD b, Bruce A. Perkins MD, MPH a, * a b
Division of Endocrinology and Metabolism, Department of Medicine, University of Toronto, Toronto, Ontario, Canada Division of Neurology, Department of Medicine, University of Toronto, Toronto, Ontario, Canada
a r t i c l e i n f o
a b s t r a c t
Article history: Received 8 September 2014 Received in revised form 24 February 2015 Accepted 24 February 2015
Objective: In vivo corneal confocal microscopy (IVCCM) has been established in cross-sectional studies as a valid measure for the identification of diabetic sensorimotor polyneuropathy (DSP). We aimed to determine the predictive validity of a baseline IVCCM measure in identifying future DSP onset in patients with type 1 diabetes. Methods: We followed 65 patients with type 1 diabetes without DSP at baseline. They were followed longitudinally for a mean of 3.50.9 years and underwent IVCCM, clinical and electrophysiologic examinations at baseline and follow up. At the end of follow up, participants were assigned as new-onset cases of DSP or as controls. Predictive validity was assessed using receiver operating characteristic curves. Results: At baseline, participants were 3415 years of age with mean diabetes duration of 1812 years. The 11 (17%) new-onset cases of DSP were similar to the 54 (83%) controls in baseline age, diabetes duration, gender, glycated hemoglobin levels and electrophysiologic parameters (p0.20). However, cases of new onset had significantly lower baseline corneal nerve fibre length (CNFL) and branch density (p0.05 for each comparison), indicating that the tests were unable to discriminate sufficiently between cases of future-onset neuropathy and controls. Discussion In this analysis of a longitudinal study designed to evaluate whether a single baseline IVCCM parameter measurement could predict future DSP incidence in a cohort of participants with type 1 diabetes and without neuropathy at baseline, we were able to confirm the predictive validity of corneal nerve morphology.
Figure 2. ROC curves for the identification of new-onset DSP by IVCCM parameters (A) and functional small-nerve fibre and clinical tests (B) for the 65 controls with type 1 diabetes without DSP at baseline. (A). The AUC for CNFL was 0.78. The optimal threshold (*) for DSP prediction by CNFL was 14.9 mm/mm2 with a sensitivity of 0.82 and a specificity of 0.69 (B). The AUC for each test was not statistically different from the line of identity, representing an AUC of 0.50. (The line of identity is depicted by the diagonal solid line.) (AUC, area under curve; DSP, diabetic sensorimotor polyneuropathy; CNFL, corneal nerve fibre length; IVCCM, in vivo corneal confocal microscopy; ROC, receiver operating characteristic.)
L.E. Lovblom et al. / Can J Diabetes 39 (2015) 390e397
395
Table 3 Quantitative results of the ROC curve analysis for the identification of new-onset DSP in the 65 participants with type 1 diabetes without DSP at baseline Test
AUC (95% CI)
p value*
Optimal threshold
Sens
Spec
PVþ
PVe
LRþ
LRe
CNFL CNFT CNBD CNFD HRV TCNS CDT LDIFLARE SNAP SNCV PNAP PNCV PNFL
0.78 0.72 0.71 0.61 0.58 0.56 0.51 0.51 0.65 0.60 0.60 0.64 0.54
0.003 0.03y 0.03y 0.25 0.43 0.53 0.89 0.90 0.14 0.30 0.31 0.14 0.71
15.4 TC
Contents lists available at ScienceDirect
Canadian Journal of Diabetes journal homepage: www.canadianjournalofdiabetes.com
Original Research
In Vivo Corneal Confocal Microscopy and Prediction of Future-Incident Neuropathy in Type 1 Diabetes: A Preliminary Longitudinal Analysis Leif E. Lovblom MSc a, y, Elise M. Halpern MSc a, y, Tong Wu BSc a, Dylan Kelly BSc a, Ausma Ahmed MD a, Genevieve Boulet MD a, Andrej Orszag MD a, Eduardo Ng MD b, Mylan Ngo RT b, Vera Bril MD b, Bruce A. Perkins MD, MPH a, * a b
Division of Endocrinology and Metabolism, Department of Medicine, University of Toronto, Toronto, Ontario, Canada Division of Neurology, Department of Medicine, University of Toronto, Toronto, Ontario, Canada
a r t i c l e i n f o
a b s t r a c t
Article history: Received 8 September 2014 Received in revised form 24 February 2015 Accepted 24 February 2015
Objective: In vivo corneal confocal microscopy (IVCCM) has been established in cross-sectional studies as a valid measure for the identification of diabetic sensorimotor polyneuropathy (DSP). We aimed to determine the predictive validity of a baseline IVCCM measure in identifying future DSP onset in patients with type 1 diabetes. Methods: We followed 65 patients with type 1 diabetes without DSP at baseline. They were followed longitudinally for a mean of 3.50.9 years and underwent IVCCM, clinical and electrophysiologic examinations at baseline and follow up. At the end of follow up, participants were assigned as new-onset cases of DSP or as controls. Predictive validity was assessed using receiver operating characteristic curves. Results: At baseline, participants were 3415 years of age with mean diabetes duration of 1812 years. The 11 (17%) new-onset cases of DSP were similar to the 54 (83%) controls in baseline age, diabetes duration, gender, glycated hemoglobin levels and electrophysiologic parameters (p0.20). However, cases of new onset had significantly lower baseline corneal nerve fibre length (CNFL) and branch density (p0.05 for each comparison), indicating that the tests were unable to discriminate sufficiently between cases of future-onset neuropathy and controls. Discussion In this analysis of a longitudinal study designed to evaluate whether a single baseline IVCCM parameter measurement could predict future DSP incidence in a cohort of participants with type 1 diabetes and without neuropathy at baseline, we were able to confirm the predictive validity of corneal nerve morphology.
Figure 2. ROC curves for the identification of new-onset DSP by IVCCM parameters (A) and functional small-nerve fibre and clinical tests (B) for the 65 controls with type 1 diabetes without DSP at baseline. (A). The AUC for CNFL was 0.78. The optimal threshold (*) for DSP prediction by CNFL was 14.9 mm/mm2 with a sensitivity of 0.82 and a specificity of 0.69 (B). The AUC for each test was not statistically different from the line of identity, representing an AUC of 0.50. (The line of identity is depicted by the diagonal solid line.) (AUC, area under curve; DSP, diabetic sensorimotor polyneuropathy; CNFL, corneal nerve fibre length; IVCCM, in vivo corneal confocal microscopy; ROC, receiver operating characteristic.)
L.E. Lovblom et al. / Can J Diabetes 39 (2015) 390e397
395
Table 3 Quantitative results of the ROC curve analysis for the identification of new-onset DSP in the 65 participants with type 1 diabetes without DSP at baseline Test
AUC (95% CI)
p value*
Optimal threshold
Sens
Spec
PVþ
PVe
LRþ
LRe
CNFL CNFT CNBD CNFD HRV TCNS CDT LDIFLARE SNAP SNCV PNAP PNCV PNFL
0.78 0.72 0.71 0.61 0.58 0.56 0.51 0.51 0.65 0.60 0.60 0.64 0.54
0.003 0.03y 0.03y 0.25 0.43 0.53 0.89 0.90 0.14 0.30 0.31 0.14 0.71
15.4 TC
