Neuropathy-specific alterations in a Mexican population of diabetic patients

This study comprised 375 non-diabetic control subjects and 243 T2DM patients, who were attending to the Family Medicine Clinics from the Mexican Institute of Social Security (IMSS).

The cross-sectional study was conducted from 2009 to 2010 in Mexico City. Among those excluded were pregnant women, patients with type-1 diabetes mellitus, kidney or heart diseases, hepatitis, peripheral arterial insufficiency diagnostic, positive HIV serology, and self-reported presence of acute infections or chronic illness such as allergies and autoimmune diseases, which could alter the circulating levels of cytokines. Also, those subjects participating in a weight reduction program were excluded. A similar population of age- and gender-matched non-diabetic, healthy subjects served as control.

We recruited patients attending the Medical Biochemistry Research Unit and outpatients attending the Neurology Service at the Hospital of Specialties of The National Medical Centre Siglo XXI. The patients were referred from family medicine clinics (numbers 5, 7, 9, 15, 16, 21, 22, 26, 38, 79 and 161), corresponding to the 03 delegation of IMSS, with a diagnosis of T2DM according to the criteria of the American Diabetes Association (ADA) or World Health Organization consultation criteria [7, 8].

A total group of 649 patients from the IMSS Family Medicine Clinics, fulfilling ADA criteria for T2DM, were screened for neuropathy, using the Michigan Neuropathy Screening Instrument (MNSI) [9], with high specificity (88.4%) and sensitivity (78.1%) according to the criteria of the consensus of San Antonio [9]. Patients with a screening value ≥2 were considered neuropathic [10], as summarized in Fig. 1. Risk factors evaluated by the Michigan Neuropathy Screening Instrument (MNSI) were: age, time of evolution of diabetes, high-density lipoproteins-cholesterol (HDL-Cholesterol) and glycosylated hemoglobin (HbA1c). MNSI can be used as a relatively simple and reliable method for clinical and epidemiological screening and assessment of asymptomatic diabetic peripheral neuropathy (ADPN), as previously reported [7], with high specificity albeit lower sensitivity compared to nerve conduction velocity tests [8].

Furthermore, patients with diabetic neuropathy were submitted to a complete neurological review to determine sensory or motor neuropathy. Sensory neuropathy was characterized by the presence of feeling of numbness, burning pain, lack of feeling in the feet, cramps, feeling of itching, pain to cover their feet with the sheet, the ability to discriminate the temperature level at their feet during bathing, the presence of wounds or ulcers, exacerbation of discomfort during the afternoon to night, their ability to feel their feet when walking, the presence of dryness or cracks on their feet or amputation, the reduction or loss of sensation to touch and/or temperature (which is symmetrical in glove and sock), unique loss of thermal sensitivity (even though this variety is very rare), reduction or loss of the vibration sense, as well as the sense of position, usually by affecting the myelinated long fibers and other proprioceptive afferent fibers.

The presence of alterations in small fibers is associated with increase in the sensation of temperature, the presence of pain, a feeling of painful burn and some symptoms related to significant autonomic alteration due to damage of unmyelinated fibers. Motor neuropathy was characterized by the presence of weakness in the extremities (usually distal), accompanied or non-by altered sensibility [11].

The group of patients with sensory neuropathy was sub-classified in a group with hypoesthesia or less sensitivity, and hyperesthesia with greater sensitivity, to identify differences between these groups. Patients with sensory polyneuropathy and hypoesthesia were characterized by decreased sensitivity to touch, vibration and pressure with monofilament; as well as reduced or absent osteotendinous reflexes. By contrast, patients with sensory polyneuropathy and hyperesthesia were characterized by increased sensitivity to touch, vibration and pressure with monofilament; as well as normal or reduced osteotendinous reflexes. A landmark for this group was the manifestation of allodynia. The neuro-conduction studies were not taken into consideration because we did not identify any atypical clinical picture, such as superimposed chronic inflammatory demyelinating polyneuropathy (CIDP), and there was no case of any familial history of neuropathy [12].

Diagnosis of cardiovascular autonomic neuropathy was documented by previous cardiology review, with tachycardia at rest, physical exercise intolerance, orthostatic hypotension and silent myocardial ischemia. Patients answered a questionnaire to evaluate alteration in esophageal motility, diagnosis of gastroparesis diabeticorum, constipation, diarrhea or fecal incontinence, and genitourinary disorders like diabetic cystopathy, erectile dysfunction, and retrograde ejaculation, also, sexual dysfunction with loss or reduction of vaginal lubrication [13]. Patients showed autonomic neuropathy if they presented the following symptoms: chronic diarrhea (predominantly at night), gustatory salivation, dizziness on rising, erectile dysfunction and orthostatic hypotension by reducing 30 or more mmHg in systolic pressure in the morning [14, 15].

The patients also answered a questionnaire with 15 questions (options yes or not). Those who answered positively for the presence of resting tachycardia, intermittent blurry vision, posture instability, syncope, postprandial hypotension and intolerance to physical exercise, were further analyzed for their electrocardiograms, deep breathing, heart rate variability and orthostatic hypotension [9, 16‐18]. Only seven patients were diagnosed with pure autonomic neuropathy (DAN) in the original group. We excluded this neuropathic group from the study because of the scarce sample size.

A complete medical history and neurological examination included cranial nerves, sensory and motor functions, as well as tendon reflexes. The vibration sense was evaluated with a tuning fork of 128 Hz, and a filament of 10 g tensiometer was used to evaluate the light touch perception. Polyneuropathy was defined as a failure to elicit the knee or Achilles muscle stretch reflexes with or without symptoms of sensory disturbance on both feet; or typical loss of sensation to pin prick and distal deep pain and/or loss of proprioception with sensory ataxia and a positive Romberg sign and/or burning sensation in the feet at night or day [14, 19].

Patients were classified after a complete physical-neurological examination, in the following groups: A control group of non-diabetic subjects (ND), a control group of diabetic non-neuropathic patients (DNN) and three test groups that included patients with sensory polyneuropathy with hypoesthesia (DS_hP) or hyperesthesia (DS_HP), as well as patients with motor neuropathy (DMN). As the total number of measurements could not be eventually assessed in all the individuals, a random selection within each group was performed.

Anthropometric indexes: Body weight was taken using a digital scale (Seca, Hamburg, Germany) to the nearest 0.1 kg and height was measured using a Seca 255 stadiometer to the nearest 0.1 cm. Body mass index was calculated as previously reported [20].

Blood pressure measurements: Systolic and diastolic blood pressures were obtained by auscultation using a sphygmomanometer (ALPK2, Tokyo, Japan) with appropriate cuff size for arm length. Four blood pressure readings were taken for each participant in the right arm in sitting position, resting 1 min between measurements, and considering the final levels of systolic and diastolic blood pressures (SP and DP, respectively) as the mean of the last three readings for each case. The standard equation was used to calculate the mean arterial pressure (MAP), as follows: MAP = DP+(SP-DP)/3 [21].

Clinical biochemistry: For biochemical analysis, blood was drawn after a 12 h fast. Glycated hemoglobin, serum glucose, total cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol and triglycerides were measured using an auto-analyzer instrument (Instrumentation Laboratory, ILAB 350, Barcelona, Spain). Plasma interleukin-6, interleukin-10, resistin, intercellular adhesion molecule-1, vascular adhesion molecule-1 and E-selectin were determined by standard enzyme-linked immunosorbent assay (ELISA) kits with inter-assay coefficients of variation <9% (R & D Systems, NE, Minneapolis, USA).

Creatinine and blood urea nitrogen were also assessed, and the Glomerular Filtration Rate (GFR) was calculated using the quadratic GFR equation as described elsewhere [22].

NGF determination: NGF concentration in plasma was measured by ELISA using the DuoSet kit (R & D Systems, NE, Minneapolis, USA), according to manufacturer instructions.

Zero values in IL-6 and IL-10 measurements were numerous in several groups of diabetic patients. The percentages of zero values were analyzed independently of non-zero values, as a simple approach to cope with data containing many null measurements [23].

Data management and analysis were performed double blind. Frequency distributions were analyzed using a Chi-squared test with Yate’s correction for continuity on 2 × 2 contingency tables. Working data were tested for normality by a Kolmogorov-Smirnov test. Most of the measurements did not fulfill normality, even after logarithmic or square root transformations for correcting normality and variance heterogeneity [24]. Comparisons among groups were performed by the non-parametric Kruskal-Wallis test (with a Dunn post test), using the statistical software GraphPad InStat version 3.00 (San Diego California USA). Datasets used for analysis are provided as a Additional file 1 (Table S1).

Different letters indicate statistical differences among groups (i.e. in five groups denoted as “a”, “b”, “c”, “ab” and “abc”, the groups “a”, “b”, “c” are statistically different from each other, the group “ab” only differs from “c” and the group “abc” is not significantly distinct from any other). Sample sizes may differ among conditions within the same group due to unavailability of individual blood sample for all the measurements. Graphics were constructed using Origin version 8 (Northampton USA).

Demographic and clinical characteristics of the study groups are summarized in Table 1. The average age was similar, except for DMN and DNN patients, with the highest and lowest values, respectively. Female gender proportion was more than a half in each group, with the highest percentage in the ND group (75%). According to a Chi-squared test, the percentages of female gender from the DS_hP (63%) and the DMN (56%) groups significantly differ from ND subjects, but there were not differences among the diabetic groups.

Other differences were found after physical examinations (Table 2); i.e., the BMI was most elevated in diabetic non-neuropathic patients and the group with sensory polyneuropathy with hypoesthesia, compared to healthy subjects. Although the percentages of previously diagnosed hypertensive patients did not differ between the groups of diabetic patients, the mean arterial pressures were significantly elevated in the neuropathic groups respect to ND subjects.

Fasting glucose levels were significantly increased in all groups of diabetic patients compared to healthy subjects (Table 2). Glycated hemoglobin (HbA1c), expressed as the percentage of total hemoglobin, was analyzed as an index of long-term insulin resistance (Table 2). Values of this parameter for patients with DS_HP and DMN were again higher than the rest of the groups, followed in decreasing order by DS_hP, DNN and ND. Hypertriglyceridemia was also observed in all the diabetic groups (Table 3). In addition, while the median total cholesterol was below 200 mg/dl in all groups, high-density lipoprotein (HDL-Cholesterol) levels were reduced in patients presenting sensory polyneuropathy with hypoesthesia. Even though the low-density lipoprotein levels (LDL-Cholesterol) were similar among groups, the DS_HP group presented the lowest average in the HDL/LDL ratio, followed by patients with DS_hN (Table 3).

We further explored some hormones and cytokines, to determine if their levels vary according to the type of neuropathy. The median circulating NGF level was reduced in more than 90% in all the groups of diabetic patients, compared to healthy controls. Among them, the DS_HP group exhibited the lowest values (Fig. 2).

The distribution of IL-6 levels exhibited a wide range of values in diabetic patients; however, the high proportion of null values in these groups hindered the statistical analysis. To approach this problem most directly, we analyzed zero and non-zero values as recommended elsewhere [23]. Indeed, the ranges of non-zero measurements in diabetic patients were significantly higher than in non-diabetic subjects (Fig. 3a). Indeed, the high percentages of null IL-6 values in diabetic patients (24% in DNN, 41% in DS_hP, 42% in DS_HP and 34% in DMN), contrasted with a very low percentage of zero values (0.8%) in the group of ND subjects (Fig. 3b).

A similar scenario occurred for IL-10, where most of the groups exhibited high percentages of zero measurements (27% in DNN, 30% in DS_hP, 35% in DS_HP and 30% in DMN), contrary to the ND group with no null measurements (Fig. 3d). The distributions of non-zero measurements were similar among all groups (Fig. 3c).

Some cell-adhesion proteins in plasma have been described as predictors of diabetic neuropathy [25]. Interestingly, all the groups of diabetic patients show increased levels in the vascular adhesion molecule-1 (VCAM) compared to healthy subjects (Table 4), where the DMN group displayed the highest value. The inter-cellular adhesion molecule-1 (ICAM) was also elevated in all diabetic groups in >70%, compared to the ND group, although no differences were found among the non-neuropathic patients and those with diabetic neuropathies (Table 4). Median levels of E-selectin also increased between 131 and 311% in diabetic patients, compared to healthy subjects (Table 4). Again, patients with DMN and DS_HN presented the highest values.

We explored kidney function, since nephropathy is a frequent complication in T2DM [3]. The glomerular filtration rate (GFR) was calculated to determine signs of renal dysfunction. The data suggest that although all the neuropathic patients show a reduced GFR, the most affected are those with motor neuropathy (Fig. 4).