ELECTROPHYSIOLOGICAL STUDIES OF NERVE FUNCTION IN PATIENTS ON CAPD OVER LONG PERIODS

The authors studied electrophyhsiological parameters in 29 patients (23 nondiabetics, six diabetics, mean age 47.9, range 16-74 yr.), who had been on CAPD for three to five years. The parameters studied were: motor nerve conduction velocities of the median, ulnar, and peroneal nerves, and sensory nerve conduction velocities of the median and ulnar nerves. These parameters were measured before or within the first month of CAPD and every six months thereafter. The changes in nerve conduction velocities were studied using simple linear regression analysis, and calculating the differences of the slopes from zero. Comparison of the mean values of the conduction velocities in non-diabetics at the beginning of CAPD and in age-matched, nonuremic patients showed significantly (p < 0.001) lower values in all parameters measured. The initial nerve conduction velocities in diabetic patients were significantly (p < 0.01) slower than in non-diabetics From the Divisions of Nephrology and Neurology, Toronto Western Hospital and Department of Medicine, University of Toronto This work was supported by the Peritoneal Dialysis Research Fund of the Toronto Western Hospital Key Wards: Neuropathy, CAPD, Complications, Uremic Neuropathy. except sensory conduction velocity of the ulnar nerve. The slopes of nerve conduction velocities plotted against time were not significantly different from zero for any nerve in both non-diabetics and diabetic patients. In conclusion, no significant change in peripheral nerve function was seen after long term CAPD in either diabetic or non-diabetic patients. Peripheral neuropathy is a frequent complication of chronic renal failure; it is seen in about 65% of patients before the start of dialysis or within a few months of its inception (I). Uremic neuropathy still remains an enigma although it has been attributed to accumulation of substances with a molecular weight of 500-5000 Daltons (middle molecules) (2). This hypothesis is based on two observations: first, the reduction of serum creatinine and urea by hemodialysis to levels corresponding to those found in mild chronic renal failure, which is usually associat ed with normal nerve conduction velocity, did not result in consistent improvement in nerve conduction velocities (3, 4 ); secondly, patients on chronic intermittent peritoneal dialysis with even higher levels of serum creatinine and BUN than those observed on hemodialysis rarely experience uremic neuropathy (5). The latter observation is in keeping with the fact that peritoneal dialysis has higher clearance of middle molecules, and a lower clearance of small molecules such as urea and creatinine than hemodialysis (6). The effect of hemodialysis on uremic neuropathy has been variable; some workers report a deterioration (3,4) and others, a stabilization (8). Contrary to the initial reports (5), our experience showed that uremic neuropathy progresses in patients on intermittent peritoneal dialysis. Initially we expected that CAPD would improve neuropathy because this modality gives a better clearance of middle molecules than does hemodialysis. However, earlier studies found that uremic neuropathy does not show any significant improvement on CAPD over short periods (10, 11). Since the changes in neuropathy occur slowly over many years, the uremic neuropathy developing in patients on CAPD over long periods may have a different cause than that which develops over short periods. This paper presents the results of sequential measurements of nerve electrophysiologi cal parameters in patients on CAPD for three years or more. PATIENTS AND METHODS We studied 23, non-diabetic patients (nine males, 14 females) with a mean age of 47 years (range 16-74) at the beginning of the study, and six diabetics (four males, two females) with a mean age of 51 years (range 40-70). Eight non-diabetics and three diabetics were studied for three years, 10 non-diabetics and three diabetics were studied for four years, and five nondiabetics were studied for five years. In 23 non-diabetic patients the renal failure was due to: chronic glomerulonephritis (9) , chronic pyelonephritis (4), polycystic kidney disease (2), nephrosclerosis (3), others (3), and unknown (2). None of our patients had a disease known to be associated with peripheral neuropathy other than uremia. The electrophysiological studies were carried out before or within the first month of CAPD and every six months thereafter. The parameters stud ied were motor nerve conduction velocities of the median, ulnar , and peroneal nerves, and sensory nerve conduction velocities of the median and ulnar nerves. The nerve conduction studies were performed according to the method described by Mayer (12), except that surface electrodes were used rather than needle electrodes. To ensure the adequate skin temperature measurements were done after patients were kept in a room temperature of 22°C for a half hour or more. For the statistical analysis, a leastsquares regression line was fitted to measurements obtained from each patient during the course of the study. The null hypothesis was that the slope of conduction velocity, as a function of time, was zero. The nerve conduction velocities of individual patients at the beginning and at the end of the study were compared using a paired t -test. The sensory nerve conduction velocity could not be recorded on one occasion or more (i.e. the nerve did not respond to the electrical stimulation) in all diabetics -39/74 recordings. Therefore, the sensory nerve conduction velocity in diabetic patients was not analyzed. In contrast, this occurred in six of 23 non-diabetic pat ients 23/292 recordings. "No response" to motor stimulation was infrequent -two of six diabetics (6/108 recordings), four of 23 non-diabetics (15/462 recordings). This was observed only with peroneal nerve recordings. If no response was recorded on more than two occasions during the follow-up period, the patient was excluded from the analysis of that particular nerve. Otherwise, we assigned the lowest value recorded in that particular patient for the nerve during the entire follow-up period. This approximation tends to underestimate deterioration during CAPD treatment (23). The mean ± S.D. of 127 agematched normal control subjects was taken to represent normal range of nerve conduction velocity.