Clinical features that distinguish PLS, upper motor neuron–dominant ALS, and typical ALS

Defining diagnostic categories for diseases of unknown cause is difficult, particularly for rare disorders such as primary lateral sclerosis (PLS).¹ Definitions are important, however, for patients who want to know their diagnosis and what the future holds, and for researchers who try to enroll a homogeneous sample of participants in clinical studies. Diagnostic boundaries allow trialists to avoid diluting the effect of a drug among different disorders and researchers to measure possible environmental and genetic risks for specific diseases. To date, there have been no clinical trials in PLS and virtually no studies examining causes for the disorder, in part because it is rare, but also because of a lack of consensus on how best to define PLS as a clinical entity.

Previously, we described clinical diagnostic boundaries for PLS and upper motor neuron dominant (UMN-D) amyotrophic lateral sclerosis (ALS) in a cohort of patients who had only upper motor neuron (UMN) signs at the time of first examination.² One group with no lower motor neuron (LMN) findings on examination or EMG 4 years after onset had significantly better prognosis and function than patients with ALS, and was defined as clinically pure PLS. A second group that had predominantly UMN signs with minor LMN findings and outcome intermediately between PLS and ALS was defined as UMN-D ALS. We selected the forth year as the time of most reliable diagnosis because most patients with only UMN signs at onset who would later develop UMN-D or typical ALS had acquired LMN signs before 4 years from onset.

Given the differences in function and prognosis between these categories, as well as the prolonged duration to diagnosis, it is essential to be able to anticipate which patients are destined to develop LMN signs, or conversely those who will retain only UMN signs indefinitely. We undertook the current project to determine how clinical features can be used to predict a diagnostic category and provide an indication of eventual outcome.

METHODS

We reviewed the records of all 24 patients (9 PLS and 15 UMN-dominant ALS) who had repeated follow-up for more than 4 years, as well as those of 10 randomly selected control patients with ALS seen in 1984–2007. The patients were identified through a search of the motor neuron disease database at the Eleanor and Lou Gehrig MDA/ALS Research Center. Eligible participants were at least age 40 and patients with PLS had negative genetic tests to minimize overlap with hereditary spastic paraparesis³; none had other identifiable cause for degeneration of the corticospinal tracts than motor neuron disease. Diagnostic categories were defined as clinically pure PLS (only UMN signs on examination, and normal EMG of three limbs, a bulbar muscle, and cervical, thoracic, and lumbosacral paraspinal muscles at least 4 years after disease onset); UMN-D ALS (predominant UMN findings on examination, with LMN signs of weakness, wasting, or fasciculation limited to one to two muscles, or EMG abnormalities limited to sparse fibrillation potentials/positive sharp waves, fasciculation potentials, or minor enlargement of motor unit potentials in one to two muscles); or ALS (meeting revised El Escorial Criteria for ALS).⁴

The primary aim of the study was to determine which clinical features at the initial visit and during follow-up distinguish between eventual diagnostic categories and provide an indication of outcome for patients with only UMN signs at symptom onset. Clinical features assessed were as follows: baseline demographic data (age, time to first evaluation, site of onset, and sex); weight loss (≥5 pounds); breathing capacity (forced vital capacity [FVC], FVC percent predicted)⁵; the ALS Functional Rating Scale (ALSFRS-R)⁶; neurologic examination of strength (manual muscle testing [MMT]) using Medical Research Council (MRC) scoring,⁷ tone, tendon reflexes, fasciculation, muscle bulk, voluntary rapid limb movements, sensory modalities (pain, temperature, light touch, vibration or position sense), and emotional lability; urinary abnormality including any of frequency, urgency, or incontinence; a report or assessment of abnormal cognition; and the N-acetylaspartate:creatine ratio in the motor cortex by magnetic resonance spectroscopy (MRS). EMG findings were reported previously.² The ALSFRS-R, MMT, and FVC were treated as continuous variables, while other examination findings were treated as categorical variables (normal vs abnormal) or in the case of rapid limb movements as normal, slow, or unable. Fasciculation was considered abnormal except in the feet. The study received institutional review board approval (Columbia University IRB # AAAB3026). Patients were exempt from signing consent because this was an analysis of existing data that was recorded by the investigators in such a manner that subjects cannot be identified.

Analysis of variance F tests for continuous variables and χ² tests for categorical variables assessed differences in baseline data among the three diagnostic categories. Linear and generalized mixed effects models were used to assess the relation between examination data and diagnostic group over time.

RESULTS

At first examination (table 1), the lowest score on the weakest muscle (p < 0.001), the site of onset (p = 0.041), and the time to first evaluation (p = 0.051) discriminated between eventual diagnostic group. Patients with PLS were stronger (mean MRC score = 4.92) than the UMN-D patients (4.38), who were stronger than the ALS patients (3.28) on the weakest muscle. An MRC score of 4 or less on any muscle was associated with the diagnosis of ALS (100%), but not PLS (0%, p = 0.0001). Patients with PLS were more likely to have limb onset than the ALS group (89% vs 80%), and the time to first evaluation was longer in the PLS group (57.7 months) than the other groups, which were similar. There were no significant differences between groups in other clinical features at the first visit, including overall MMT score, the side (left v right) of onset, and MRS assessment of the UMN.

Across all visits (tables 2 and 3), overall muscle strength (p = 0.003), ALSFRS-R (p = 0.009), and FVC (p = 0.026) were significant predictors of group assignment when controlled for age at onset. In all three analyses, there was a decreasing trend in the score with age. The average MMT score differed among the three groups in the arms (p < 0.001), legs (p = 0.001), and bulbar muscles (p < 0.001). The largest difference in ALSFRS-R occurred between the PLS and ALS groups; the decline of the ALSFRS-R score was nonlinear in all three groups. For FVC, as for the ALSFRS and MMT, patients with PLS had the highest overall score, followed by those with UMN-D and then ALS across time. UMN-D and ALS groups were more likely to have visible muscle atrophy (UMN-D = 67%, ALS = 90%, PLS = 22%, p = 0.009) or to have lost weight (UMN-D = 60%, ALS = 70%, PLS = 0%, p = 0.004), even when controlled for dysphagia in the analyses (p = 0.021), than the PLS group, and patients with ALS were more likely to have hyporeflexia (ALS = 70%, UMN-D = 7%, PLS = 11%, p = 0.001) than the other groups. The presence of bulbar findings could be used to discriminate between PLS and ALS groups (atrophy/fasciculation/dysphagia), PLS and UMN-D groups (dysarthria/dysphagia), and UMN-D and ALS groups (fasciculation). Spasticity was more likely to occur in the UMN-D than the ALS group (100% vs 70%, p = 0.03). There were no significant differences between groups in other features over time, including urinary or sensory changes.

DISCUSSION

Several studies found that clinical PLS, while rare, has a slower rate of progression and better overall prognosis than ALS.^2,8-11 One group reported that patients with PLS and patients with ALS differed in the frequency of muscle atrophy and bulbar signs, as well as the symptom of stiffness,¹¹ while some series included patients with only UMN signs and also those with minor LMN signs or EMG denervation.^9-11

We believe that the differentiation of patients with and without limited LMN signs is important, however. The exceptionally long course and high level of functioning of some patients with PLS seems to apply only to those who have no LMN signs and a normal EMG. Our patients with UMN-dominant signs who developed LMN signs or EMG abnormalities, no matter how subtle, had a shorter survival than patients with clinically pure PLS and their level of function declined to a range similar to those with typical ALS² (appendix e-1 on the Neurology® Web site at www.neurology.org).

The time to diagnosis for the disorder is also disputed, again perhaps because of lack of discrimination between patients who have only UMN signs and those with subtle LMN signs. In our cohort, 77% of patients with only UMN signs at first evaluation but who later developed the sparse LMN signs of UMN-D ALS or more marked findings of ALS did so by the 4th year after onset. EMG abnormalities appeared after a median of 3.17 years and LMN signs appeared 6 months later, with 35% of patients being affected between years 3 and 4.² We selected the 4th year after symptom onset as the most reliable time for the diagnosis of PLS, knowing that some patients will still develop LMN signs even after that time.¹² One group concurred with this time frame and offered modified diagnostic criteria,¹³ while another maintained that 3 years from onset is sufficient to diagnose their patients with PLS,¹¹ again lumping those that we would separate as PLS and UMN-D ALS.

Given the different prognosis and level of function of the three groups, pure PLS, UMN-D ALS, and typical ALS, in conjunction with the prolonged time to diagnosis, it is crucial to be able to anticipate which patients are destined for which diagnostic group before that time. Our data indicate that PLS can begin with pseudobulbar signs, but that it is unusual and should warn a clinician that LMN signs are likely to develop in the future. Those destined to develop LMN signs were more likely to have bulbar onset than the PLS group, and over time, the prominence of bulbar signs differed in all three groups. The time to the first evaluation, a gauge of the rate of disease progression, was longer in the patients with PLS than the patients with UMN-D or ALS. Patients with PLS can also develop weakness, but it is usually mild and often generalized in an UMN distribution. Focal weakness of MRC grade 4 or less at the initial visit was invariably due to LMN disease and should inform the clinician that LMN signs along with a worse prognosis are present. Similarly, the absence of muscle atrophy separated the PLS group from both the ALS and UMN-dominant groups. Weight loss was also associated with the presence of LMN signs. Even when controlled for dysphagia, patients with ALS and patients with UMN-D ALS were significantly more likely to have lost weight than were patients with PLS, a finding in agreement with data that ALS, or perhaps LMN dysfunction, is associated with metabolic abnormalities.¹⁴ Reduced FVC was also seen in our UMN-D and ALS groups, but not the PLS group, suggesting that FVC measures LMN function in the phrenic nerve more than the UMN.

Regarding differentiation of the three diagnoses, the PLS group was distinguished from the UMN-D group by less weight loss, muscle weakness, and atrophy, as well as fewer pseudobulbar signs. The PLS group also differed from the ALS group in site of onset, fewer bulbar signs, and less weight loss, weakness, atrophy, and hyporeflexia. The UMN-D and ALS groups differed in frequency of spasticity, as well as hyporeflexia, bulbar signs, and weakness in the patients with ALS.

Certain clinical signs are helpful in discriminating among eventual diagnostic groups. The emerging picture of clinically pure PLS is characterized by slow progression, high function, normal breathing, and no weight loss, and includes no evidence of LMN dysfunction on examination or EMG 4 years after symptom onset. Prior to the forth year, we counsel patients about the potential outcomes and the importance of periodic examinations for evidence of LMN dysfunction.

A weakness of the study is the small sample size, yielding low power and the potential for bias. It is also not possible to determine whether clinical differences result from differences in pathogenesis. Larger prospective longitudinal studies are needed to confirm our findings as well as to determine overall outcome. Postmortem examination of patients with clinically pure PLS and UMN-D ALS is needed to search for pathologic differences, including the pathognomonic findings in ALS of LMN degeneration along with ubiquitinated TDP-43^15-17 and Bunina body^18–20 inclusions.

In the meantime, we suggest that patients with bulbar onset, a muscle weaker than MRC grade 4, weight loss, reduced FVC, meaningful loss of function on ALSFRS-R, or EMG changes not be reassured of the more benign diagnosis of PLS, and that patients with only UMN signs have periodic assessments of all of these measures because a change in any can signal the development of LMN disease. Reassurance can be given to patients and a homogeneous sample could be enrolled in a trial before year 4 who are normal in all of the above areas, knowing that the firmest diagnosis cannot be made until that time.

AUTHOR CONTRIBUTIONS

B.C. performed the statistical analyses.