There have been no prospective, population-based studies of diabetic amyotrophy and mononeuropathies in subjects with diabetes. However, some prevalence figures for these types of neuropathy can be derived from a few cross-sectional studies. In a cross-sectional survey based in Rochester, Minnesota, asymptomatic carpal tunnel syndrome (CTS) was found in 22% of those with type 1 diabetes and 29% of those with type 2 diabetes, whereas the corresponding prevalence for symptomatic cases was 11% and 6%, respectively (40). Ulnar and femoral cutaneous entrapment was found in 2% of type 1 diabetes and 1% of type 2 diabetes subjects. Cranial mononeuropathy and trun-cal radiculopathy were not observed in the Rochester population, whereas proximal asymmetric polyneuropathy was identified in 1% of type 1 diabetes and type 2 diabetes subjects (40). No incidence data were available for any of these types of neuropathy.
In a cross-sectional, hospital-based study, O'Hare et al. (41) studied the presence of various types of neuropathy (by interview assessment) in 800 consecutive subjects with diabetes (336 type 1, 464 type 2) treated at one diabetes center and 100 subjects without diabetes attending an otolaryngology clinic. The prevalence of neuropathy in subjects with diabetes was 22.9%. Less common types included amyotrophy (total prevalence: 0.8%), oculomotor neuropathy (0.1%), peroneal neuropathy (0.1%), and truncal neuropathy (0.1%). Risk factors for neuropathy in type 1 diabetes were age (56.7 ± 15 years in subjects with neuropathy vs 44.9 ± 18 years in those without neuropathy, p < 0.001) and duration of diabetes (17 ± 10 years in subjects with neuropathy vs 13 ± 9 years in those without neuropathy, p < 0.001). In type 2 diabetes, age was also associated with neuropathy (64.2 ± 9 years in subjects with neuropathy vs 60 ± 12 years in those without neuropathy, p < 0.002). Unlike previous epidemiological studies, the O'Hare study documented the presence of various types of neuropathy, in addition to diabetic peripheral neuropathy, in a large group of subjects with diabetes. However, the interview assessment of sensory neuropathy was likely to be insensitive, and the study was neither prospective nor population-based.
Another cross-sectional study assessed the presence of median mononeuropathy (MM) in 414 subjects with diabetes and mild neuropathy enrolled in a randomized controlled treatment trial of tolrestat vs placebo (42). MM was defined by criteria from nerve conduction studies and was differentiated from CTS, a diagnosis based on history, physical exam, and electrophysiological findings. The prevalence of MM was 23%, and it was associated with longer duration of diabetes in both type 1 (22.5 years in subjects with MM vs 16 in those without MM, p = 0.003) and type 2 diabetes (8.8 years in subjects with MM vs 7.0 years in those without MM, p = 0.034). In type 2 diabetes, height and body mass index were also predictors of MM. The inclusion of subjects with mild neuropathy already enrolled in a treatment trial creates selection bias. A fourth cross-sectional study of CTS included 470 subjects from a diabetes clinic, a neurology clinic, and community volunteers (43). Fifty-two had neither diabetes nor neuropathy, 81 had diabetes without neuropathy, and 337 had diabetes and neuropathy. The prevalence of CTS, determined by clinical evaluation using accepted criteria, was 2% in subjects without diabetes, 14% in patient with diabetes without neuropathy, and 30% in subjects with diabetes and polyneuropathy. CTS was linked to a longer duration of diabetes (14 ± 12.5 years in those with neuropathy vs 10.8 ± 10.7 years in those without neuropathy). This study was one of few that examined the point prevalence of CTS in subjects with diabetes, and like the one previously described, suggests that MM and CTS are common in people with diabetes, especially in those with diabetic peripheral neuropathy. However, since these studies are neither prospective nor population-based, these results may be biased and cannot be extrapolated to the general diabetic population.
Case-control studies have addressed whether diabetes is a significant risk factor for CTS. In the earliest study, 156 cases were identified from the population by self-report (n = 28) and from neurology clinic (n = 128), and controls were 476 subjects without CTS from the community (44). The diagnosis of CTS was confirmed by neu-rophysiological testing. Significant independent risk factors were height (per 1 cm) (OR = 0.9), frequent wrist activities (OR = 1.1), leg varicosities in men (OR = 9.8), menopause in the previous year (OR = 2.3), and hysterectomy (OR = 1.8). In order to examine the effect of selection bias, risk factor analysis was repeated for the 28 cases derived from the population (rather than clinic) and yielded similar results. The low diabetes prevalence in cases (2.6%) and controls (3.7%) resulted in lower power and precluded any conclusion about diabetes as a risk factor. Three other case-control studies concluded that diabetes is a risk factor for CTS. In one large retrospective case-control study of enrollees of New Jersey Medicare or Medicaid programs during a 3-year period, 627 people who underwent open or endoscopic CTS procedures were selected as cases, and 3740 controls were frequency-matched by age and gender with the cases (45). Risk factors were inflammatory arthritis (OR 3.1, 95% CI 2.2-4.2), corticosteroid use (OR 1.6, 95% CI 1.2-2.1, DM (OR 1.4, 95% CI 1.2-1.8), female sex (OR 1.6, 95% CI 1.3-2), and hemodialysis (OR 9, 95% CI 4.2-19.6). Although this study did have a large sample size, it suffers from several limitations: retrospective study design; selection bias (subjects who underwent surgery may have had more severe CTS); potentially inaccurate case ascertainment (clinical conditions were defined by diagnostic codes and prescriptions, not by chart review); and a failure to adjust for obesity, which may causally influence the association between diabetes and CTS.
Two more recent case-control studies also suggest that diabetes may be risk factor for CTS but again have limited validity because of similar weaknesses in study design to those already described (46,47). In the first, both cases and controls were drawn from hospitals and clinics, and controls included subjects with upper limb symptoms who did not meet electrophysiological criteria for CTS (46). In the second, cases were subjects who had undergone a CTS procedure, and controls were seen for general reconstructive surgery or presented with acute hand symptoms to a plastic surgeon (47). Of note, one major strength of the first study was that it did present the results of both multivariable analysis (in which diabetes was a significant risk factor with CTS) and stratified analysis using obesity as a stratifying factor (46). After adjusting for obesity, diabetes was no longer a significant risk factor.
CHRONIC INFLAMMATORY DEMYLINATING POLYRADICULOPATHY (CIDP) IN PERSONS WITH DIABETES
CIDP is a relatively new diagnosis. In 1991, the American Academy of Neurology defined diagnostic clinical and electrophysiological criteria for CIDP (48). All studies on CIDP are cross-sectional and clinic-based. In one study, cases were requested by letter from neurologists in four Thames health regions in southeast England (49). The personal case series of two investigators were also included. The subjects' clinical data were reviewed to confirm the diagnosis of CIDP by standard diagnostic criteria established by an ad hoc subcommittee. The degree of certainty of the diagnosis was classified as definite, probable, possible, or suggestive. Population statistics were obtained from the Office of Population Censuses and Survey. The prevalence of definite and probable CIDP in the Southeast Thames region was 1/100,000. In this study, case ascertainment may have been affected by the fact that reported cases were excluded if there was no available confirmatory data. In addition, authors noted that a tertiary referral center in one of the four Thames health regions had a special focus on inflammatory neuropathy (49). In a second similar study, cases and relevant clinical data were requested from all 94 neurologists in New South Wales, and 84% responded. Population data was derived from the Australian national census (50). The crude prevalence of CIDP in New South Wales was 1.9/100,000. And the crude annual incidence was 0.15/100,000 (50).
One large cross-sectional study examined the possible association between diabetes and CIDP. Among 1127 subjects seen in an electrophysiology lab over a 14 month period, 189 subjects (16.8%) had diabetes (51). The prevalence of CIDP (diagnostic criteria modified from the American Academy of Neurology in 1991) was 16.9% in subjects with diabetes and 1.8% in subjects without diabetes (OR 11.04, 95% CI 6.1-21.8, p < 0.001). There was no difference in the prevalence of CIDP in type 1 and type 2 diabetes (26.7% vs 16.1%, p = 0.49). This study demonstrated selection bias because subjects referred to the electrophysiological laboratory tended to have more severe neuropathy and possibly a greater likelihood of CIDP. In order to determine which clinical manifestations are linked to types of neuropathy other than diabetic polyneuropathy, Lozeron et al. (52) examined and performed nerve conduction studies on 100 consecutive subjects with diabetes with symptomatic neuropathy referred to one neurology center over 3.25 years. The prevalence of CIDP (by clinical presentation and nerve conduction studies) was 9%. The sample size of this case series was small, and the results reflect selection bias and cannot be generalized to the whole diabetic population.
Was this article helpful?