A multicenter, randomized, double-blinded trial of pyridostigmine in postpolio syndrome

Methods.

Results.

A total of 126 patients fulfilled study criteria and were enrolled in the trial between February 1996 and April 1997. Sixty-four patients were randomly assigned to pyridostigmine, and 62 to placebo. The two groups were similar with respect to baseline demographic and clinical characteristics and baseline outcome values (table 1), except that a somewhat greater proportion of patients assigned to pyridostigmine were women (42/64 [65.6%] versus 34/62 [54.8%]), and serum IGF-I was somewhat higher in patients assigned to placebo. All patients completed the trial (figure).

• Download figure
• Open in new tab
• Download powerpoint

Figure. N refers to number of patients who came in for evaluations. Not all patients completed all outcome evaluations at each visit. The percentage of patients with less than 80% compliance included 11/64 (17%) pyridostigmine-treated patients and 7/62 (11%) placebo-treated patients for whom there was missing information.

In general, the study drug was well tolerated. Four severe adverse events (three in pyridostigmine and one in placebo-treated patients) and 12 reportable adverse events in 11 patients (nine pyridostigmine and two placebo-treated patients) were noted during the study. The severe adverse events are described in table 2. Among the patients with reportable adverse events, two patients (pyridostigmine-treated) discontinued the medication. This occurred due to painful muscle and gastrointestinal cramps in one patient, and nausea, diarrhea, vomiting, and faintness in another patient. One patient, assigned to placebo, decreased the medication for the remainder of the study due to the sensation of feeling drugged and blurred vision. The medication was stopped and then restarted at the full dose, using an incremental initiation schedule in five patients due to reportable adverse events. This occurred in four pyridostigmine-treated patients due to the symptoms of painful muscle cramps, severe abdominal pain (possibly urinary tract infection), mild continuous nausea and diarrhea, and continued profuse sweating and chest pain (each symptom reported in one patient), and in one placebo-treated patient due to nausea and diarrhea (probable viral gastroenteritis). The following reportable adverse events that occurred in three pyridostigmine-treated patients did not result in change in medication dose: fractured fibula, fractured rib while operating machine, and herpes zoster twice in one patient. The most common adverse events reported at the time of the treatment assessments were gastrointestinal. Thirty-five of 64 (55%) patients in the pyridostigmine group and 12/62 (19%) patients in the placebo group noted diarrhea and loose stool (p < 0.001). Eighteen of 64 (28%) patients in the pyridostigmine group and 8 of 62 (13%) in the placebo group noted nausea and vomiting, or gastrointestinal upset (p = 0.035).

Compliance with the study medication was good, and similar in the pyridostigmine and placebo-treated patients (see the figure). Forty-five of 64 (70%) pyridostigmine and 45/62 (73%) placebo-treated patients used at least 80% of the expected number of pills of medication. The percentage of patients with less than 80% compliance included 11/64 (17%) pyridostigmine-treated patients and 7 of 62 (11%) placebo-treated patients for whom there was missing information on compliance. Five of 64 (8%) pyridostigmine and 1 of 62 (2%) placebo-treated patients stopped the medication due to severe and reportable adverse effects. In summary, 43/64 (67%) pyridostigmine-treated patients and 42/62 (68%) placebo-treated patients were at least 80% compliant with the medication and were within the correct time window at the 6-month treatment visit.

The effectiveness of blinding was analyzed. Sixty-seven percent of patients (41/61) in the pyridostigmine-treated group and 33% of patients (20/61) in the placebo-treated group guessed that they were receiving the active treatment. A chi-square test comparing the proportions of patients guessing that they were receiving active treatment among the two treatment groups provided evidence for unblinding (p = 0.001).

In the primary intent to treat analysis, the study showed no difference in the pyridostigmine and placebo-treated patients with regard to health-related quality of life, most measures of isometric strength, and subjective fatigue (as assessed by two fatigue scales) at 6 months of treatment (table 3). In addition, no significant difference was observed in serum IGF-I levels between patient groups at 6 months of treatment. However, the very weak muscles, those that were 1 to 25% of predicted normal at baseline, showed a trend to increasing strength with pyridostigmine over time, and at 6 months, a greater increase in strength in pyridostigmine-treated patients than placebo-treated patients. The percent difference from baseline in pyridostigmine-treated patients was 10.5 ± 66.8% (n = 33) at 6 weeks, 20.4 ± 68.5% (n = 33) at 10 weeks, and 41.8 ± 108.5% (n = 36) at 6 months. The percent difference from baseline in placebo-treated patients was 7.9% ± 47.2% (n = 29) at 6 months (difference between groups at 6 months 33.9%, 95% CI of difference −9.5 to 73.3, p = 0.10). Approximately 10% of muscles studied were included in this group of muscles. There were no meaningful differences in pyridostigmine and placebo-treated patients in any outcome measures at 6 and 10 weeks of treatment.

A restricted analysis performed on patients who were at least 80% compliant with the medication and who were evaluated within the time window at 6 months showed similar results. In addition, a trend to increased IGF-I levels with pyridostigmine at 6 months of treatment was noted in this patient population. IGF-I levels increased from baseline by a mean of 13.1 ± 60.6 ng/mL (n = 40) in pyridostigmine-treated patients and decreased from baseline by a mean of 6.1 ± 53.9 ng/mL (n = 37) in placebo-treated patients (difference between groups 19.2 ng/mL, 95% CI of difference −6.4 to 44.8 ng/mL, p = 0.15).

Analysis of patient reports of symptoms on the interval history showed that at 6 months of treatment significantly fewer patients in the pyridostigmine group than in the placebo group noted increased weakness since the last study visit (15/63 [23.8%] versus 25/60 [41.7%], p = 0.035). There were no differences with regard to this symptom at 6 and 10 weeks of treatment. No differences were noted between groups in the observation of improved strength since the last study visit. At the time of the 6-week visit, significantly fewer patients noted muscle fatigue in the pyridostigmine group than in the placebo group (47/60 [78.3%] versus 56/61 [91.8%], p = 0.037), but not at 10 weeks and 6 months of treatment. There were no differences between groups with regard to the observation of improved muscle fatigue.

The influence of study center, gender, and selected baseline outcome values on the primary outcome (SF-36 physical functioning scale), isometric strength, and serum IGF-I levels was evaluated using an analysis of covariance model. Study center, gender, and baseline levels of the primary outcome and fatigue did not influence study results with regard to the primary outcome. Similar results were observed for isometric strength and serum IGF-I levels. A trend to increased strength in the very weak muscles with pyridostigmine was again noted after adjustment for study center, gender, and baseline fatigue and strength (p = 0.11).

Discussion.

The results of this randomized, multicenter clinical trial of a 6-month course of pyridostigmine in PPS demonstrate no statistically significant effect of the medication on health-related quality of life, isometric strength, subjective fatigue, and serum IGF-I levels. However, a nonsignificant improvement of isometric strength with pyridostigmine was seen in the very weak muscles at 6 months of treatment in both unadjusted and adjusted analyses. In addition, in the subgroup of patients compliant to their treatment assignment, there was a suggestion of an increase in serum IGF-I in pyridostigmine-treated patients relative to placebo-treated patients. In the exploratory analyses, at 6 months of treatment, patients who were assigned to pyridostigmine were significantly less likely to complain of increased muscle weakness (as per standardized interval medical history). Despite some positive trends from secondary and tertiary analyses, this study was negative and the possible benefits of the medication require further study.

The finding of no statistically significant effects of pyridostigmine in PPS on the primary and secondary outcomes of the study did not reflect the previous experience of some physicians and patients that there appeared to be a clinical benefit with the medication in a proportion of patients.16,17 Several explanations for this discrepancy may hold. First, pyridostigmine may actually have no significant clinical benefit in PPS as seen in this clinical trial. Because the preliminary studies with pyridostigmine were small and only one was placebo-controlled, the initial positive results observed may have been misleading. Alternatively, this study could be falsely negative for a number of reasons. First, the outcome measures may not have been responsive to a treatment effect. The measures chosen were those that had undergone some evaluation in PPS and would be most practical for the study. PPS is a relatively newly recognized disorder, and this was the first multicenter therapeutic trial in PPS. Therefore, further work to develop the most appropriate outcome measures for PPS clinical trials remains to be done. The SF-36, which was the primary outcome for the study, is a generic, nondisease-specific measure of health-related quality of life. It was chosen for the study because it is well-accepted, widely used, and appeared to be assessing some important constructs in PPS. However, a recent study35 showed that in ALS, which is usually a very rapidly progressive neuromuscular disease, the SF-36 correlated only with major changes in patient status (e.g., progression from economic dependence stage to caregiver dependence stage). It is therefore possible that the physical functioning scale of the SF-36 was not responsive enough to a change in strength and that another instrument developed specifically for PPS may have produced a different result. Isometric strength was chosen as an objective measure of muscle function in the study primarily because this measure had been well evaluated and standardized in PPS, and equipment for the measure was already available at all participating centers. However, based on the known physiologic actions of pyridostigmine, measures of muscle fatigue or endurance may have been more likely to show a response with pyridostigmine than isometric strength. Second, physical activity was not assessed during the trial, and some investigators and patients noted that physical activity increased during the trial, but that fatigue levels remained stable. Therefore, an objective measure of physical activity in parallel with fatigue may have been more appropriate. Third, the timing of medication dose before assessment of isometric strength was not standardized, and this could have influenced the results of this measure by increasing the variance. The variance of the isometric strength evaluations was high (table 3), possibly due to clinical heterogeneity or to timing of medication, if effective. Fourth, the actual dose of pyridostigmine used may not have been adequate. The dose used for the study was one that had appeared to be adequate in several small preliminary clinical trials,16-18 but a dose ranging study had not been performed before initiation of the trial. The dose of pyridostigmine used in myasthenia gravis can be higher (rarely, up to 1 gram per day) than used in this study. In addition, compliance with the study treatment was approximately 70% in both groups, reducing the chance of finding a medication effect. Fifth, it is possible that pyridostigmine is useful for PPS only when combined with other interventions, such as judicious exercise, which may be useful in certain PPS patients.13

A trend to increased strength with pyridostigmine was observed at the 6-month treatment assessment in the very weak muscle groups (1 to 25% predicted normal). In addition, in the very weak muscle group a trend over time for increasing strength with pyridostigmine was observed. These results should be interpreted with caution because they come from secondary analyses with multiple testing. They are, however, compatible with the known action of the medication, and consistent with the results of preliminary studies of pyridostigmine on strength in PPS.18,36 Pyridostigmine would be most likely to acutely improve strength in those muscles in which NMJ transmission is most unstable. The weakest muscles would be expected to have the largest motor units (due to denervation and compensatory reinnervation),6 with resultant greatest physiologic stress on overly extended motor axonal arborizations and greatest deficits in NMJ transmission.37 However, the trend to increased strength was observed only at 6 months, and not earlier, suggesting that the drug’s possible action on muscle strength may be due to a chronic, trophic effect and not an acute effect mediated by a cholinesterase-induced amelioration in NMJ transmission. An acute effect of the medication may not have been captured because of the lack of objective measures of muscle fatigue and endurance in the study.

There are many difficulties encountered in performing a clinical trial of pyridostigmine in PPS.38 The patients vary greatly in the degree of initial motor neuron involvement at acute polio with great heterogeneity even in different limbs of individual patients, in the degree of recovery, and in the type and severity of current symptoms. This is reflected in the great variability of baseline characteristics of our patient population (see table 1). We attempted to reduce some of this variability by entering only ambulatory patients into the study. In addition, the diagnostic criteria for PPS have not yet been strictly defined, although several proposals have been made.39,40 We utilized a modified version of these proposals. The importance of the symptoms reported on patient function and quality of life has not been well assessed. Therefore, it is not clear which would be the most important symptom to utilize as the primary outcome for a clinical trial. In addition, PPS is a slowly progressive disease, and any treatment would require a sufficient period of time to show efficacy, compared to placebo. In our trial, little change was observed in most outcome measures over the 6-month period of the trial. There was evidence for unblinding among patients in this study. This may influence the interpretation of the results of subjective patient symptom reports. Given the subjective nature of the primary outcome, if the study were positive, unblinding could be an explanation for a falsely positive result. Because the study was negative, this is not an issue for the primary outcome but would need to be addressed in any future clinical trials with pyridostigmine. Despite these difficulties, we launched the trial owing to the pathophysiologic rationale for pyridostigmine in PPS, the promising results from preliminary trials of pyridostigmine in PPS, and the lack of a specific treatment for this illness.

This trial did not show a clear benefit of pyridostigmine. Another pyridostigmine study using different outcome measures and evaluating the possibility of increased strength in weak muscles at 6 months might yield a different result.