Abstract
Background: Although peripheral neuropathy (PN) occurs after bariatric surgery (BS), a causal association has not been established.
Objectives: To ascertain whether PN occurs more frequently following BS vs another abdominal surgery, to characterize the clinical patterns of PN, to identify risk factors for PN, and to assess if nerve biopsy provides pathophysiologic insight.
Methods: Retrospective review identified patients with PN after BS. The frequency of PN was compared with that of an age- and gender-matched, retrospectively evaluated cohort of obese patients undergoing cholecystectomy.
Results: Of 435 patients who had BS, 71 (16%) developed PN. Patients developed PN more often after BS than after cholecystectomy (4/126; 3%) (p < 0.001). The clinical patterns of PN were polyneuropathy (n = 27), mononeuropathy (n = 39), and radiculoplexus neuropathy (n = 5). Risk factors included rate and absolute amount of weight loss, prolonged gastrointestinal symptoms, not attending a nutritional clinic after BS, reduced serum albumin and transferrin after BS, postoperative surgical complications requiring hospitalization, and having jejunoileal bypass. Most risk factors were associated with the polyneuropathy group. Sural nerve biopsies showed prominent axonal degeneration and perivascular inflammation.
Conclusions: Peripheral neuropathy (PN) occurs more frequently after bariatric surgery (BS) than after another abdominal surgery. The three clinical patterns of PN after BS are sensory-predominant polyneuropathy, mononeuropathy, and radiculoplexus neuropathy. Malnutrition may be the most important risk factor, and patients should attend nutritional clinics. Inflammation and altered immunity may play a role in the pathogenesis, but further study is needed.
Obesity is of epidemic proportions and an increasing health problem in the United States and the Western world. The prevalence of obesity surveyed between 1988 to 1994 was estimated at 20% in US men and 25% in US women.1 Its prevalence has increased significantly in children, adolescents, and adults of both genders in the United States over the last two decades.2 Obesity is associated with an increased prevalence and incidence of heart disease, stroke, obstructive sleep apnea, type 2 (insulin-resistant) diabetes mellitus, dyslipidemia, hypertension, hepatobiliary disease, cancer, endocrine disorders, psychosocial disturbances, and orthopedic complications.3
No dietary approach has achieved uniform, long-term success for the morbidly obese.4 In 1991, an NIH Consensus Conference concluded that bariatric surgery (BS) was an effective treatment option for morbidly obese patients (body mass index [BMI] ≥ 40 kg/m2)5 who “are well-informed and motivated” and that the surgery should be done in a facility having experienced surgeons and other health care professionals with appropriate follow-up.6 The recommended criteria for surgical intervention included a BMI of >40 kg/m2 or a BMI of >35 kg/m2 with accompanying co-morbidity. Gastric bypass and gastric restrictive procedures were considered acceptable operations for morbid obesity.7 Jejunoileal bypass was not recommended4 because of its association with serious postoperative complications.7 Currently, there are two primary, NIH-approved operative approaches: vertical banded gastroplasty (GP) and Roux-en-Y gastric bypass (GB).8 BS procedures for obesity are currently increasingly performed and have been the topic of discussion in mass media and widely adopted by people from all walks of life including celebrities.
Well-reported neurologic complications of BS are peripheral neuropathy (PN), burning feet syndrome, meralgia paresthetica, myotonic syndrome, myelopathy, Wernicke–Korsakoff encephalopathy, and lumbosacral plexopathy.9–12⇓⇓⇓ PN has been reported to improve with vitamin supplementation,12 thus raising the possibility that PN after gastric procedures has a nutritional cause. Thiamine deficiency has been postulated to cause PN after BS.13 However, these reports were not based on controlled studies. In this study, we compare the frequency of PN after BS with the frequency of PN after another abdominal surgery, describe the clinical patterns of PN, and identify risk factors.
Designs/methods.
Using Mayo Clinic Health Science Research coded diagnoses lists, we identified the medical records of patients coded for obesity and BS at Mayo Clinic in Rochester, MN, from January 1, 1985, to December 31, 2001. Based on detailed examination of all medical records, we identified the patients who developed PN. We also identified and reviewed 300 medical records of age- and sex-matched patients who were obese and had undergone open cholecystectomy at Mayo Clinic, Rochester, MN, during the same time period. These medical records were individually reviewed using the same procedures, criteria, and categorization of neuropathy as were used in BS patients. In both cohorts, PN was defined by clinical criteria (characteristic clinical symptoms and signs [weakness, sensory symptoms and deficits, reflex loss, etc.] and confirmatory electrodiagnostic tests that localized the problem to the peripheral nerves). We included all types of PN such as length-dependent symmetric polyneuropathy, mononeuropathy, multiple mononeuropathy, radiculoplexus neuropathy, polyradiculoneuropathy, small-fiber neuropathy, large-fiber neuropathy, and so on. Neither group was surveyed for subclinical neuropathy. We specifically excluded patients who had neuropathy preoperatively.
A patient was defined as having PN after BS if he or she had one or more gastrointestinal operations for morbid obesity (BS) and developed symptomatic, clinically defined PN after BS. A patient was defined as having PN after cholecystectomy by the same criteria. The following exclusion criteria were used: 1) BS patients who had an operation(s) for an indication other than morbid obesity or cholecystectomy patients who had an operation(s) for a reason other than gallstones or cholecystitis; 2) PN developed after operation but from another known cause(s) such as alcohol abuse, heavy metal intoxication, monoclonal gammopathy, associated necrotizing vasculitis, or toxic exposure, or the patient had pre-existing PN preoperatively; 3) essential data were missing from medical records; 4) conditions that affect weight, such as metastatic cancer, untreated hyperthyroidism, pregnancy, or other; 5) neurologic conditions that would interfere with assessment of PN symptoms, such as multiple sclerosis, severe cervical or lumbosacral radiculopathy, stroke, or other.
Surgery.
All BS patients underwent one or more of four types of BS: GB, GP, pancreaticobiliary bypass, and jejunoileal bypass (JIB). All cholecystectomy patients underwent open cholecystectomy. Patients with a laparoscopic approach to BS or cholecystectomy were excluded. Postoperative gastrointestinal symptoms (nausea and vomiting, dumping syndrome, and diarrhea) were considered prolonged if symptoms occurred on a regular basis for a period of >3 months after surgery.
Weight measurement.
Weight (kg), height (m), and BMI (kg/m2) were recorded at the time of BS or cholecystectomy and on follow-up visits. In patients with PN, the postsurgery weight, height, and BMI were taken from the visit closest to the onset of the symptoms of PN. In patients without PN, postsurgery weight, height, and BMI were recorded at the most recent follow-up visit. A patient was defined as having reached maximal weight loss when there was no more than 10% change in weight during several consecutive follow-up visits (i.e., weight had plateaued).
Neuropathic evaluations.
In the case of polyneuropathy and radiculoplexus neuropathy, the characteristics and distribution of the neuropathy were quantitated using the Neuropathy Impairment Score (NIS), a standard composite score of weakness, reflex loss, and sensation loss.14 We defined the course of PN as acute if the symptoms reached maximum in 1 month, as subacute if the symptoms reached maximum between 1 and 3 months, and as chronic if the symptoms reached maximum in >3 months after BS.
Laboratory methods.
Laboratory data included hemoglobin, total lymphocyte count, and serum concentrations of vitamin B12, folate, thyroid-stimulating hormone, cholesterol, triglyceride, albumin, total iron-binding capacity, fasting blood glucose, and glycosylated hemoglobin. Serum concentrations of thiamine and other nutritional markers such as vitamin A, D, E, or K were tested only occasionally. To assess the role of inflammatory and immune causes of neuropathy, other tests were performed, such as erythrocyte sedimentation rate, rheumatoid factor, antinuclear antibody, and serum protein electrophoresis.
Electrophysiologic methods.
Conventional nerve conduction and needle electromyography were done using Nicolet Viking (Madison, WI) instruments and the laboratory normal values.
Quantitative sensory, autonomic, and thermoregulatory sweat testing methods.
Quantitative autonomic testing was performed using an autonomic reflex screening test.15 Quantitative sensory testing was carried out by computer-assisted sensation evaluation system IV (CASE IV, WR Electronics, Stillwater, MN ).16 Thermoregulatory sweat test was performed by a modification of the Guttmann quinizarin sweat test.17
Histologic methods.
Paraffin sections were stained with hematoxylin and eosin, Masson trichrome, Luxol fast blue/periodic acid Schiff, Congo red, methyl violet, and Turnbull blue. Immunohistochemistry preparations were obtained for leukocytes (CD45) and for macrophages (CD68). Semithin epoxy sections were stained with methylene blue. Teased fibers were prepared and graded by previously defined pathologic criteria.18
Nutritional management post surgery.
Most patients who had BS performed at Mayo Clinic attended a nutritional clinic run by endocrinologists who specialize in managing obese patients. Issues attended to included dietary prescription, goal setting for weight loss, lifestyle changes, metabolic and nutritional evaluation, and supplementation as well as management of existing co-morbidities. Although obese, the patients having cholecystectomy did not routinely attend nutritional clinics.
Statistical analysis.
Data are shown using descriptive statistics for clinical description of PN. Categorical data were compared using Fisher exact or χ2 test; continuous data are presented as mean values ± SD and were compared using Wilcoxon rank sum test for two-group comparison and analysis of variance for three-group comparison (with post-hoc Bonferroni test). A value of p < 0.05 was used for significance.
Results.
The BS group consisted of 435 patients. The cholecystectomy control group consisted of 300 age- and gender- matched cholecystectomy patients. Of the latter group, only 123 had open cholecystectomy for gallstone or cholecystitis, the others having been excluded because they had cholecystectomy at the time of liver transplantation, malignancy, or other extensive abdominal surgery. Seventy-one (16%) of the BS patients developed PN, whereas only 4 (3%) of the cholecystectomy patients developed PN (p < 0.001). In the BS group, the following patients were excluded: Fourteen developed PN from another known cause, 4 had severe systemic diseases that can cause weight loss, 46 had other neurologic diseases that affected limb function, and 15 had missing data regarding the surgery type. From the cohort of BS patients who did not develop PN and did not have exclusion criteria, we identified 87 who had frequent follow-up visits with weight and laboratory measurements to serve as a BS control group without neuropathy.
Clinical characteristics.
BS patients.
Table 1 provides characteristics of the BS patients with PN (n = 71) and without PN (n = 87). No significant differences between mean age, preoperative weight, postoperative weight, or preoperative BMI between the two groups were noted. Both BS patients with and without PN lost weight after surgery (mean 43 ± 24 kg with PN and 33 ± 23 kg without PN; p < 0.001 in each). Although no significant difference in preoperative and postoperative weights between the groups was found, BS patients with PN lost more weight than BS patients without PN postoperatively (p = 0.02) and postoperative BMI was lower (p = 0.03). BS patients with PN reached maximum weight loss faster than BS patients without PN (mean 8 ± 4 vs 19 ± 13 months; p < 0.001). It is unlikely that the patients without PN after BS developed neuropathy at a later time, as they were followed for a longer period than the time it took to develop neuropathy in the PN after BS group (80 ± 76 vs 61 ± 71 months; p = 0.01). In patients with and without PN after BS, the most frequently performed surgical procedures were GB (n = 47 with PN and n = 62 without PN) and GP (n = 16 with PN and n = 23 without PN). In general, there were no differences between the types of BS procedures performed in both groups except for higher frequency of JIB in BS with PN (n = 9 vs 1; p = 0.005).
Table 1 Characteristics of bariatric surgery patients with and without neuropathy
Open cholecystectomy patients.
These patients (n = 123) did not have significant weight loss after surgery. Compared with the BS group, the cholecystectomy group was older (mean age 60 vs 50 years; p < 0.001), had fewer women (47 vs 80%; p < 0.001), was less obese before surgery (mean presurgery weight 99 vs 139 kg; p < 0.001), had less weight loss after surgery (mean weight loss 0.5 vs 37 kg; p < 0.001), and was followed longer (mean follow-up time 83 vs 76 months; p < 0.0001). There was no difference in the rate of diabetes mellitus between the two cohorts (28% of cholecystectomy patients and 30% of BS patients [p = 0.5] were diabetic). As the cholecystectomy patients as a group did not lose weight, the time to maximum weight loss could not be determined.
Neuropathic characteristics of PN after BS.
There were three distinct clinical patterns of PN that developed after BS: sensory-predominant polyneuropathy, mononeuropathy, and radiculoplexus neuropathy. Table 2 provides characteristic data on these three groups of patients.
Table 2 Characteristics of three types of neuropathy after bariatric surgery
Sensory-predominant polyneuropathy.
All patients (n = 27) had symmetric sensory symptoms and signs, and some had distal motor weakness involving hands and feet. The clinical course was known for 25 patients. Most patients (n = 18) had an insidious onset and chronic course, but five had a subacute onset and two had an acute onset. The initial symptoms, known in 23 patients, were sensory in 18 patients and included dead-type numbness (n = 10), pain (n = 11), and other positive sensory symptoms such as paresthesia (n = 4) or a feeling of a tight band in the feet (n = 1) (some patients had more than one type of sensory symptoms). Weakness of the feet or legs was the initial symptom in five patients. Three patients had sensory symptoms starting in the hands. Pain eventually developed in 16 patients. The types of pain included aching (n = 5), sharp or stabbing (n = 6), and burning (n = 5). Sensory symptoms involved the head and trunk in four patients. Over time, weakness developed in nine patients and was distal and lower limb predominant. Foot drop developed in four patients. Cramping was found in three patients. Autonomic symptoms, observed in 12 patients, included constipation (n = 2), lightheadedness or syncope (n = 6), impotence (n = 2), and urinary incontinence or frequency (n = 3). Cranial nerve symptoms observed in three patients included double vision and difficulty swallowing.
Neurologic examination showed sensory deficits or motor weakness predominantly in the distal lower extremities. Deep tendon reflexes were reduced in 18 patients and normal in 9 patients. Cranial nerve examination was normal in all (despite the presence of cranial nerve symptoms in three patients). The mean NIS was 21 (range 0 to 74). Commonly, patients had sensory symptoms consistent with small-fiber neuropathy, three of whom had normal neurologic examinations. Three of six patients had increased protein concentrations in the CSF (>45 mg/dL); the mean CSF protein for the six patients was 68 mg/dL (range 27 to 195 mg/dL). The CSF white blood cell count was normal in all. Sedimentation rates were increased in 5 of 28 patients for whom they were performed. There were three patients who had positive antinuclear antibodies and one patient with positive rheumatoid factor.
Electrophysiologic studies showed moderate to severe large-fiber PN in 11 patients, mild PN in 8, and normal studies in 6 (those patients with normal studies clinically had small-fiber neuropathies). There was no electrophysiologic evidence of demyelination. Seven patients had quantitative autonomic testing, with two showing mild to moderate degree of abnormalities. Thermoregulatory sweat tests performed in four patients were abnormal in all, showing regions of anhidrosis involving both distal limbs and truncal areas. Quantitative sensory testing performed in four patients was abnormal in all, showing variable degrees of sensory involvement affecting all fiber classes.
Mononeuropathy.
In this group (n = 39), there were 31 patients with median neuropathy at the wrist (carpal tunnel syndrome), 1 patient with radial neuropathy (who also had median neuropathy at the wrist), 1 patient with superficial radial sensory neuropathy, 2 patients with ulnar neuropathy at the elbow, 1 patient with greater occipital neuropathy, 2 patients with peroneal neuropathy at the fibula head, 1 patient with lateral femoral cutaneous neuropathy, and 1 patient with sciatic neuropathy. The mononeuropathy developed acutely in 12 patients, subacutely in 8, and insidiously in 17. The patients with lateral femoral cutaneous, sciatic, radial motor, superficial radial sensory, greater occipital, and peroneal neuropathies all had acute onsets. The mononeuropathies had asymmetric involvement (only one limb) in 22 patients. Seventeen patients had bilateral mononeuropathies, which all were median mononeuropathies at the wrists, except for one patient with bilateral ulnar neuropathies at the elbows.
Radiculoplexus neuropathy.
Three patients had lumbosacral radiculoplexus neuropathy (lower limb syndrome), and two had cervical radiculoplexus neuropathy (upper limb syndrome). The symptoms began asymmetrically and remained unilateral in four patients and became bilateral in one with lumbosacral radiculoplexus neuropathy. The onset was acute for three patients and subacute for two patients. The first symptom was severe pain in four patients and numbness in one. Eventually, all patients developed pain. The types of pain were described as burning or shooting (n = 2), aching (n = 2), and excessive sensitivity to normal touch (contact allodynia) (n = 2). Weakness of the involved limb followed shortly after the onset of pain. Foot drop was present in three patients and wrist drop in one. Autonomic, cranial nerve, truncal, or CNS symptoms or signs were not seen in this group. CSF analysis was normal in the two patients evaluated. The sedimentation rate was normal in three patients. The mean NIS was 35 (range 5 to 59) in the radiculoplexus neuropathies. One patient had quantitative autonomic testing that showed mild to moderate degree of autonomic abnormalities. Electromyography and nerve conduction studies were consistent with axonal disorders involving the root, plexus, and nerves. Quantitative sensory testing performed in two patients showed variable degrees of sensory involvement affecting different fiber classes.
Neuropathic characteristics of PN after cholecystectomy.
All four patients with PN after cholecystectomy had mononeuropathies (all had median mononeuropathy at the wrist [carpal tunnel syndrome]). Mononeuropathies were still more frequent in the BS group than the cholecystectomy group (p < 0.01). Sensory-predominant polyneuropathy and radiculoplexus neuropathy were not identified in the cholecystectomy control group. There were no differences in frequency of serum concentrations of vitamin B12 or folate postoperatively between the two groups.
Pathologic alterations.
Five patients (four polyneuropathy and one radiculoplexus neuropathy) had sural nerve biopsies. Prominent, active axonal degeneration was the main neuropathic abnormality. The rate of axonal degeneration in teased nerve fibers (figure 1) was markedly increased in all nerves (mean 56%, range 25 to 84%). This finding implies a recent active pathologic process. Axonal degeneration was confirmed on paraffin and epoxy preparations that showed many degenerating profiles (see figure 1). One biopsy had multifocal fiber degeneration. An abnormal degree of mononuclear inflammatory cells involving the epineurium (figure 2) and the endoneurium (figure 3) was seen in all patients. In three biopsies, small collections (10 to 50 cells) of perivascular inflammatory mononuclear cells were present, whereas in the other two biopsies, the perivascular inflammatory mononuclear cell collections were of moderate size (50 to 100 cells). Hemosiderin-laden macrophages were present in two biopsies. No definite evidence of vasculitis was seen, although the inflammatory cells involved vessel walls in three cases, and in one there was neovascularization associated with hemosiderin (see figure 2). Four of the five patients biopsied had acute or subacute onset of the neuropathy, which occurred within 3 months after BS (in keeping with the active axonal degeneration observed).
Figure 1. Closely aligned teased nerve fibers fixed in glutaraldehyde and osmium tetroxide (top) and transverse semithin epoxy section stained with methylene blue (bottom) from sural nerves of patients with peripheral neuropathy after bariatric surgery. The top panel shows numerous myelin ovoids along the lengths of most teased nerve fibers, indicative of active axonal degeneration. The observation that all the fibers are at the same early stage of axonal degeneration implies a common insult to these nerve fibers. In the lower panel, most of the myelinated fibers (especially large fibers) are also undergoing active axonal degeneration (degenerating fibers appear as dark blotches). These findings demonstrate a severe pathologic process causing active axonal degeneration.
Figure 2. Longitudinal paraffin sections stained with hematoxylin and eosin (top) and Luxol fast blue/periodic acid–Schiff (bottom) from sural nerves of patients with peripheral neuropathy after bariatric surgery. The top panel shows a moderate-size collection of mononuclear inflammatory cells surrounding epineurial blood vessels. The bottom panel shows a scarred epineurial blood vessel surrounded by new vessel formation (neovascularization) and hemosiderin-laden macrophages (arrowhead). These findings demonstrate an immune process and are associated with active axonal degeneration (as demonstrated in figure 1).
Figure 3. Serial longitudinal paraffin sections from a sural nerve of a patient with peripheral neuropathy after bariatric surgery, stained with Luxol fast blue/periodic acid–Schiff (top), reacted to lymphocytes (CD45) (middle), and reacted to macrophages (CD68) (bottom). These sections demonstrate a mononuclear cell infiltrate surrounding an endoneurial microvessel. The bottom panel confirms that many myelinated fibers are undergoing axonal degeneration as other breakdown products are being taken up by macrophages throughout the endoneurium. The endoneurial mononuclear cell infiltrates suggest that an inflammatory or immune process is involved in the pathogenesis of the neuropathy.
Risk factors for PN after BS.
The risk factors associated with PN after BS are listed in table 1 and included a greater absolute weight loss, a faster rate of weight loss, a lower postsurgery BMI, lower serum albumin and transferrin concentrations, having surgery performed outside of Mayo Clinic, having the JIB procedure, prolonged postoperative gastrointestinal symptoms (nausea and vomiting, diarrhea, and dumping syndrome), presence of a postoperative complication(s) requiring rehospitalization, less vitamin and calcium supplementation, and not attending nutritional clinics postoperatively. The following factors had no significant association with development of PN after BS: age, gender, height, BMI at baseline, change in BMI, maximum weight loss, duration of hospitalization, number of operations, presence of diabetes mellitus, pre- and postoperative levels of hemoglobin concentration, total lymphocyte count, and serum concentrations of vitamin B12 folate, thyroid-stimulating hormone, cholesterol, triglyceride, blood glucose, and glycosylated hemoglobin.
A subgroup analysis of the different types of PN showed that the risk factors associated with development of PN after BS were more commonly associated with the polyneuropathy group than with the other two groups (table 3). The risk factors in the polyneuropathy group included lower serum albumin and transferrin concentrations, prolonged nausea and vomiting, not attending nutritional clinics, and presence of gastrointestinal complications. Patients with mononeuropathy (carpal tunnel syndrome) and radiculoplexus neuropathies were significantly more likely to have diabetes mellitus than patients with polyneuropathy.
Table 3 Subgroup analysis of factors associated with peripheral neuropathy after bariatric surgery according to type of peripheral neuropathy
Discussion.
Information on neurologic complications, specifically PN, after BS is limited, and to date most studies are small and not controlled. One study that addressed neurologic complications (including PN) after gastric restrictive procedures identified 16 patients with PN (3.2%) among 500 patients.9 In our series, the frequency was higher as we identified 71 (16%) patients with PN of 435 patients with BS.
The finding that the incidence of PN after BS is significantly higher than after cholecystectomy supports our hypothesis that neurologic complications of BS are related specifically to the bypass procedure itself or more likely to a state of nutritional deficiency rather than to having undergone an abdominal operation. This hypothesis is further supported by the significant association of risk factors indicating nutritional deficiency with the development of PN after BS.
A potential problem exists, however, in the use of the cholecystectomy patients as a control group. It is likely, given the “multidisciplinary” approach to bariatric patients, that they were somewhat more intensively followed postoperatively than were the cholecystectomy patients; this unequal surveillance might allow better identification of PN in the BS cohort. Given the retrospective design of the study, there was no way to entirely eliminate this bias and still have a control group. We tried to minimize this problem by choosing as our control group obese patients who had their cholecystectomy performed at Mayo Clinic and who had their ongoing follow-up care there. In fact, the cholecystectomy cohort was followed for a significantly longer time after surgery than was the BS cohort. Furthermore, surgeons and internists were the physicians doing the majority of patient care for both cohorts, and neurologists were not involved until after neurologic symptoms started. Consequently, it seems likely that the identification of neurologic problems would be similar, if not entirely equal, between the two groups. In spite of its limitations, we believe that the cholecystectomy control group was still very helpful; it permitted us to show that BS itself, and not an abdominal surgery, was the risk factor responsible for development of PN.
We identified three characteristic PN syndromes that occur after BS: sensory-predominant polyneuropathy, mononeuropathy, and radiculoplexus neuropathy. Further analysis showed that polyneuropathy and mononeuropathy were significantly more frequent in the BS group than in the cholecystectomy group. Although not statistically different from the cholecystectomy group, radiculoplexus neuropathy was found only in the BS group.
In our study, the association between risk factors and PN indicates that nutritional deficiency may play the most important role in the pathogenesis of PN after BS. Although our study is not prospective, we believe our findings are conclusive and plausible. Nausea and vomiting have been reported previously to be associated with PN after BS.9,11–13⇓⇓⇓ On average, weight loss after BS usually reaches its maximum within 12 to 24 months after surgery.6,19,20⇓⇓ In our study, excessively rapid weight loss was a risk factor associated with PN and reached its maximum in the PN after BS group in 8 months. Rapid weight loss after BS has been associated with low serum concentrations of vitamin B12, thiamine, and folate, and vitamin supplementation can significantly improve the nutritional status of these patients.21,22⇓ These deficiencies may predispose patients after BS to develop PN. Other nutritional deficiencies may also play a role.
The etiology of the PN after BS is probably multifactorial and different in each subgroup of PN. We found that most of the risk factors were significantly more associated with sensory-predominant polyneuropathies (see table 3) than with mononeuropathies or radiculoplexus neuropathies. Consequently, the nutritional deficiencies are probably most important in the pathogenesis of sensory-predominant polyneuropathy after BS. In contrast, having diabetes mellitus was more important in developing mononeuropathies and radiculoplexus neuropathies (see table 3).
In the mononeuropathy group, the etiology may not be different from usual mononeuropathies in which mechanical compression and entrapment appear to play a role. In our study, median neuropathy at the wrist (carpal tunnel syndrome) was the most common mononeuropathy after BS. To our knowledge, there have been no previous reports of carpal tunnel syndrome developing after rapid weight loss or after BS. On the contrary, carpal tunnel syndrome has been shown to have a positive correlation with increased BMI and obesity.23–25⇓⇓ It is therefore surprising to find this association. Perhaps rapid weight loss makes the nerves more susceptible to compression. In some patients, injury from mechanical retractors, patient positioning of the upper extremities, or the use of radial or ulnar intra-arterial catheters at the wrist during surgery may cause mononeuropathies, but these cases are probably rare.26 Peroneal neuropathy and lateral femoral cutaneous neuropathy (as we found in some patients after BS) have been reported to occur after weight loss (due to loss of subcutaneous tissue), although in the latter situation, it is also a common occurrence in obese people.26–28⇓⇓ Other mononeuropathies we observed (radial, ulnar, greater occipital, and sciatic neuropathy) have not previously been reported after BS or weight loss.
In the group that developed radiculoplexus neuropathy after BS, the clinical features were similar to those described in other forms of diabetic and nondiabetic radiculoplexus neuropathy.29–31⇓⇓ One patient with lumbosacral plexopathy has been previously reported after BS.10 The pathogenesis of diabetic and nondiabetic radiculoplexus neuropathies is likely due to ischemic injury from microvasculitis.29,31,32⇓⇓ One of our patients with lumbosacral radiculoplexus neuropathy after BS had inflammatory infiltrates on nerve biopsy but did not have microvasculitis. Cervical radiculoplexus neuropathy also has been shown to have inflammatory lesions on brachial plexus and upper limb nerve biopsies.33,34⇓ In our patients, we suspect the pathogenesis may have also been inflammatory or immune, but we did not biopsy any upper limb nerves. In the diabetic and nondiabetic radiculoplexus neuropathies, there is often a large associated weight loss.29,32⇓ We have suspected that this weight loss is secondary to the vasculitis and immune disorders. However, given the observation that rapid weight loss and cachexia occur in patients with BS before the onset of the radiculoplexus neuropathy, the role of cachexia in inducing an immune or microvasculitic neuropathy needs to be considered in all types of radiculoplexus neuropathies.
The role of inflammatory and immune mechanisms in PN after BS deserves further consideration. It was a striking finding that all five sural nerve biopsies had evidence of an active neurogenic disorder with prominent axonal degeneration and inflammatory infiltrates in the epineurium and endoneurium. Although none was diagnostic of vasculitis, the degree of inflammation was distinctly abnormal. These biopsies came from patients with polyneuropathy or radiculoplexus neuropathy, most of whom had acute or subacute onset of their symptoms. Consequently, the inflammatory infiltrates may not be representative of the more insidious cases of PN after BS. Furthermore, we do not believe that these pathologic findings diminish the importance of nutritional deficiencies and cachexia in the pathophysiology of the neuropathy. Rather, we believe that cachexia and nutritional deficiencies may, in fact, induce immune mechanisms of neuropathy. Although an immune hypothesis is intriguing, further work needs to be done to confirm our suspicion.
Is there a specific nutritional deficiency that causes PN after BS? We have not identified one specific nutritional deficiency; rather, the cause seems to be multifactorial nutritional deficiencies, and patients should be evaluated on a case-by-case basis. In some patients, deficiency of multiple nutritional elements may be involved. Prolonged starvation causes polyneuropathy, in some patients attributed directly to deficiency of thiamine or niacin.35 In our patients, thiamine levels were not checked routinely. Nevertheless, we observed no patients with Wernicke encephalopathy. Perhaps surprisingly, neuropathy from vitamin B12 or folate deficiency has not been reported after BS. Patients having undergone JIB and GB have been reported to have vitamin B12 or folate deficiency but not neuropathy.36,37⇓ In our study, serum concentrations of vitamin B12 and folate were not different between PN and no-PN groups in both the pre- and the postoperative periods. In addition, the number of patients with low vitamin B12 and folate was not different between the two groups, indicative that vitamin B12 and folate probably do not play an important role in development of PN after BS for the overall group (although they may for individual patients). The other vitamin deficiency states that are known to cause PN include vitamin B6 (pyridoxine) and vitamin E (tocopherol) deficiency.38 Assessment of these vitamins was not performed routinely in our patients. Copper deficiency has also recently been shown to be associated with myelopathy and sensory neuropathy often in patients who have previously had gastrointestinal surgeries.39 None of our patients had myelopathy, and copper levels were not performed.
The observation that significantly fewer patients who had their BS performed at Mayo Clinic developed PN than those who had BS elsewhere may be interpreted in several ways. It may demonstrate the good surgical technique at our institution, or, more likely, it may mean that patients who had their BS elsewhere and developed PN sought medical help from our tertiary referral institution. We believe a likely explanation is that at our multidisciplinary clinic, most BS patients attended a focused nutritional clinic, and much of the malnutrition seen in non-Mayo patients after BS was avoided in Mayo patients after BS due to nutritional supplementation. In fact, BS patients who attended nutritional clinics developed neuropathies significantly less commonly than those who did not (see table 1). This observation underscores the importance of attending nutritional clinics during the postoperative period. An endocrinologist with expertise in BS patients can appropriately manage the weight loss and the nutritional needs. If there are any complications of BS, they will more likely be detected in a timely manner and managed appropriately.
Acknowledgments
Supported in part by grants obtained from the National Institute of Neurological Diseases and Stroke (NINDS 36797).
The authors thank Peter J. Dyck, MD, for editorial help.
- Received January 26, 2004.
- Accepted May 20, 2004.
References
Letters: Rapid online correspondence
REQUIREMENTS
You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
