Early Detection of Small-Fiber Neuropathy in Diabetes

A laser-induced pain somatosensory-evoked potentials and pupillometric study

  1. Giuseppe Pozzessere, MD1,
  2. Paolo Rossi, MD1,
  3. Annarita Gabriele, MD2,
  4. Rosalba Cipriani, MD2,
  5. Antonino Morocutti, MD1,
  6. Umberto Di Mario, PHD2 and
  7. Susanna Morano, MD2
  1. 1Institute of Neurology, University “La Sapienza,” Rome, Italy
  2. 2Department of Clinical Sciences, Endocrinology, University “La Sapienza,” Rome, Italy

    Over the last 15 years, many authors have pointed out that small nerve fibers may be selectively damaged in the early stages of diabetes, leading to an early impairment of pain and temperature sensations and to a decline in autonomic nervous function (1,2).

    The CO2 pain-induced laser somatosensory-evoked potential (pSEP) evaluation may represent a new quantitative and objective approach to evaluate pain and temperature sensation functions (38). Recently, Agostino et al. (9) reported small-fiber dysfunction assessed by pSEPs in diabetic patients with different degrees of peripheral nerve damage.

    A total of 27 diabetic patients (12 type 1 [group 1] and 15 type 2 [group 2]) and 27 age-, sex-, and height-matched control subjects were included in the study. In group 1 (6 men and 6 women), the mean age was 33.6 ± 9 years, the duration of diabetes was 13.5 ± 6.7 years, and the glycated hemoglobin was 6.1 ± 1.3%; two patients had nephropathy and three had background retinopathy. In group 2 (7 men and 8 women), the mean age was 55.3 ± 8.8 years, the duration of diabetes was 8.2 ± 6.2 years, and the glycated hemoglobin was 6.8 ± 1.5%; three patients had nephropathy and four had background retinopathy. The patients reported no history of autonomic and somatic neuropathy and had negative clinical examinations (evaluated using the Neuropathy Symptom Score [NSS], the Neurological Symptom Profile [NSP], and the Neurologic Disability Score [NDS]) (10); normal nerve conduction velocity and P100 latency at visual-evoked potentials (VEPs) recording; no severe diabetic retinopathy or other opthalmological diseases; no recurrent ketoacidosis, ketonuria, or hypoglycemia; no psychiatric disorders or alcohol consumption; no cognitive impairment; and no other diseases or treatment with medications known to influence nervous system function. Two control groups (groups 3 and 4) were included in the study. Group 3 (6 men and 6 women, aged 34.1 ± 8 years) comprised healthy subjects age-, sex-, and height-matched with patient group 1, and group 4 (7 men and 8 women, aged 55.5 ± 8.7 years) comprised healthy subjects age-, sex-, and height-matched with patient group 2. All subjects were informed about the aim of the study and gave their consent. The experimental design of the study was approved by the ethics committee of “la Sapienza” University of Rome. Diabetic- and control-subject findings were compared with those obtained in our laboratories in a large number (60) of normal subjects acting as a normal reference control group.

    A portable CO2 laser stimulator (Neurolas, Florence, Italy) was used in this study for the pSEP evaluation (11,12). The most consistent and prominent components of the response to laser stimulation are seen as a negative-positive complex (Fig. 1). Positive potentials, P340 in the hand pSEPs and P400 in the foot pSEPs, were evaluated in all subjects. The upper limit of the normal range of P340 and P400 was set to mean ±2.5 SD of the values obtained in the reference group. For the pupillometric monocular evaluation (ISCAN; sample rate 50 Hz), dark diameter (DD) for sympathetic function, and pupillary constriction latency (PCL) for parasympathetic autonomic function, respectively, were measured (1316). Measurements were made in the right arm and left leg using a standard surface-stimulating and recording technique, maintaining skin temperature at 32°C. Motor nerve conduction velocity, sensory nerve conduction velocity, and sensory and motor action potential amplitude results were considered normal if the values were not >2.5 SD different from the data obtained in the reference group. In the VEP evaluation (13), P100 latency was considered abnormal if the value exceeded the mean ±2.5 SD of the control population. A blood sample to measure HbA1c levels was collected from each fasting patient in the morning (HPLC; Menarini, Firenze, Italy; upper limit of the normal range 6%). Blood glucose levels measured before each session were >120<150 mg/dl, and no hypoglycemic episodes were recorded during the neurophysiological assessment.

    The Student’s t test, the Wilcoxon’s rank-sum test, and the Fisher’s exact test were used to study differences between groups. Univariate and multivariate logistic regression analyses were used to evaluate the correlation between parameters. Moreover, to measure the extent of agreement between pSEP and pupillary autonomic function tests, the McNemar test was computed. Neurophysiological data from diabetic patients and control subjects are shown in Table 1. The P340 latency was not significantly different between diabetic patients and control subjects. One patient in group 1 and one patient in group 2 showed a bilateral increase of P340 latency above the mean ±2.5 SD of normal reference values (maximal P340 value 384 m/sec). Mean peak latency of P400 was significantly prolonged in diabetic patients compared with control subjects. Individual analysis showed abnormally prolonged P400 values in 3 of 12 patients in group 1 and in 4 of 14 patients in group 2 (maximal P400 value 438 m/sec). Abnormalities were bilateral in each case. The correlation analysis indicated that duration of disease was independently associated with P400 latency in both type 1 (r = 0.63, P < 0.05) and type 2 (r = 0.56, P < 0.05) diabetic patients. DD was significantly reduced in both diabetic groups (group 1 and 2) compared with control groups. Four patients in group 1 and five patients in group 2 presented abnormally reduced bilateral DD values. No significant PCL difference was found between diabetic patients and control subjects. A prolonged latency was found bilaterally in two patients of group 1 and in two patients of group 2. Duration of disease was independently correlated with DD in group 1 (r = −0.55, P < 0.05) and in group 2 (r = −0.53, P < 0.05). Among the patients in group 1, 58.3% presented one or more abnormalities in the neurophysiological recordings. One patient showed abnormality of P400, DD, and PCL, and two patients presented only abnormal pSEPs (in one both P340 and P400 were prolonged). No significant correlation was found between pSEPs and pupillometric parameters. One or more abnormal value of pSEPs and pupillometric parameters were found in 57.2% of group-2 patients: one patient presented abnormal values of P340, P400, and PCL and two patients showed abnormal DD and P400 results. No significant correlation was found between pSEPs and pupillometric parameters. Age, sex, and hypoglycemic treatment were not significantly different in patients with and without neurophysiological abnormalities. An early, subclinical, and selective damage of small nerve fibers, regarding both autonomic and somatic functions, in diabetic patients without clinical sign of diabetic neuropathy and electrophysiological evidence of large-fiber dysfunction, has been demonstrated in our study. We found a prolonged P400 latency in both type 1 and type 2 diabetic patients, whereas P340 latency was not significantly affected. Our results are partially contrasting with those obtained by other studies (9), but pSEP amplitude was not considered in our study. P400 latency was bilaterally abnormal in 25% of type 1 diabetic patients and 28.5% of type 2 diabetic patients, respectively.

    These data indicate the presence of a selective subclinical symmetrical hypoalgesia involving first the longest pathways of the lower limbs. This pattern of small nerve fiber dysfunction may represent the result of a length-related degeneration of peripheral fibers. In our study, this hypothesis is also supported by the pupillometric results showing a predominant dysfunction of the longest sympathetic fibers. As for quantitative sensory testing findings, the latency of cortical-evoked response reflects both central nervous system (CNS) and peripheral nervous system function. Previous studies assessing central somatosensory pathways by means of the conventional electric pSEPs have shown that the CNS may be affected with a low frequency at an early stage of diabetic disease, before the appearance of overt neurological complications (17,18). Thus, in our study, the possibility of some CNS involvement cannot be dismissed. Interestingly, in electrical pSEP studies, the CNS parameter abnormalities were frequently monolateral (17). In our patients the abnormalities were always bilateral. This symmetry of pathological findings suggests a subclinical peripheral rather than a central neurological dysfunction.

    A further issue to be discussed regards the type of cortical responses recorded after laser stimulation, since the extent to which cortical pSEPs reflect sensory or cognitive processing of nociceptive inputs is still under debate (6,19). At present, data from previous studies support the notions that cortical pSEPs 1) are unlikely to represent a purely endogenous potential and 2) measure an exogenous potential, exploring the function of pain and temperature sensitive pathways (6,7,20). In our study, we have carefully kept all factors influencing pSEPs constant; moreover, we used a standardized distraction task to separate the exogenous component of the late responses. We also found a prolongation of P400 latency in the presence of normal P340 values. These data do not reflect a cognitive dysfunction, the effect of which should not have been limited to lower limb stimulation, and are consistent with the conclusion that abnormalities of pSEP latency reflect a small-fiber dysfunction. Investigating the parasympathetic and sympathetic pupillary function we found in both type 1 and type 2 diabetic patients, a significantly reduced DD with no significant difference for latency was found. These data suggest that sympathetic dysfunction precedes parasympathetic damage, leading to a diminished size in darkness of pupils normally responding to light stimuli. No significant correlation was found between pupillometric parameters and pSEP latencies. These data indicate that somatic and autonomic nerve fibers are differently affected in preclinical stages of diabetes and are consistent with the view that cranial autonomic fibers and somatic peripheral fibers may not be damaged at the same time, confirming the necessity of a simultaneous neurophysiological assessment of different nerve fibers (21). Additionally, these data suggest that abnormal pupillary parameters may be considered a simple and useful marker of subclinical diabetic neuropathy, relatively independent of other neurological abnormalities. Duration of diabetes was significantly correlated with both autonomic and somatic parameters, whereas we failed to find any correlation between pupillometric parameters, pSEP latency, and metabolic control.

    In conclusion, this study evaluated for the first time the impairment of pain sensation by measuring pain pSEP latency in diabetic patients with no clinical or electrophysiological evidence of large nerve-fiber dysfunction. The abnormalities of pSEP latencies mainly affected the lower limbs resembling a length-related pattern of neuropathy. Pupillary study showed a contemporary but unrelated damage of small sympathetic autonomic fibers. These findings confirm that small fibers may be selectively involved and more prone to damage in diabetic patients, strengthening the necessity for an accurate, quantitative, and noninvasive simultaneous assessment of different nerve fibers. In this view, the cortical pain pSEPs may be used to evaluate the electrophysiological integrity of A-Δ fibers in diabetic patients, because no other accurate objective examinations are available to study the impairment of nociceptive sensitivity.

    Figure 1—

    Main positive component (arrows) of the pSEPs after foot stimulation (P400 wave) in a 52-year-old normal subject (A) and in an age- and sex-matched diabetic subject (B). Double traces represent averages over blocks of 30 artifact-free responses.

    Table 1—

    Neurophysiological data


    • Address correspondence to Dr. Susanna Morano, Endocrinologia Dipartimento di Scienze Cliniche, Clinica Medica 2, Università “La Sapienza,” Policlinico Umberto I, Viale del Policlinico, 00161 Rome, Italy. E-mail: susanna.morano{at}uniroma1.it.


    | Table of Contents

    Navigate This Article