Nonmotor symptoms as presenting complaints in Parkinson's disease: A clinicopathological study


  • Sean S. O'Sullivan MRCPI,

    1. Reta Lila Weston Institute of Neurological Studies, Institute of Neurology, University College London, London, United Kingdom
    Search for more papers by this author
  • David R. Williams PhD,

    1. Reta Lila Weston Institute of Neurological Studies, Institute of Neurology, University College London, London, United Kingdom
    2. Faculty of Medicine (Neuroscience), Monash University (Alfred Hospital Campus), Melbourne, Australia
    Search for more papers by this author
  • David A. Gallagher MRCP,

    1. Department of Clinical Neurosciences, Royal Free and University College Medical School, London, United Kingdom
    Search for more papers by this author
  • Luke A. Massey MRCP,

    1. Reta Lila Weston Institute of Neurological Studies, Institute of Neurology, University College London, London, United Kingdom
    Search for more papers by this author
  • Laura Silveira-Moriyama MD,

    1. Reta Lila Weston Institute of Neurological Studies, Institute of Neurology, University College London, London, United Kingdom
    Search for more papers by this author
  • Andrew J. Lees FRCP

    Corresponding author
    1. Reta Lila Weston Institute of Neurological Studies, Institute of Neurology, University College London, London, United Kingdom
    • Reta Lila Weston Institute of Neurological Studies, UCL, 1 Wakefield Street, London, WC1N 1PJ
    Search for more papers by this author


Nonmotor symptoms (NMS) are increasingly recognized as a significant cause of morbidity in later stages of Parkinson's disease (PD). Prodromal NMS are also a well recognized component of the clinical picture in some patients but the prevalence of NMS as presenting complaints, and their impact on clinical management, in pathologically-proven cases of PD is unknown. The presenting complaints of 433 cases of pathologically-proven PD archived at the Queen Square Brain Bank for Neurological Diseases were identified from the clinical case notes. 91/433 (21%) of patients with PD presented with NMS of which the most frequent were pain (15%), urinary dysfunction (3.9%), anxiety, or depression (2.5%). Presenting with NMS is associated with a delayed diagnosis of PD (Mann-Whitney U, P = 0.001). These patients were more likely to be misdiagnosed initially and were more likely to have been referred to orthopedeic surgeons or rheumatologists than neurologists (nonmotor group 5.5% vs. motor group 44.2%, χ2P < 0.0001). NMS are commonly seen as presenting complaints in pathologically confirmed PD, and initial misdiagnosis may be associated with potentially inappropriate medical interventions. Presenting with NMS does not affect the motor response to medication, but is associated with shorter disease duration (χ2P = 0.016). © 2007 Movement Disorder Society

Parkinson's disease (PD) is a common neurodegenerative disorder characterized by bradykinesia, rigidity, and rest tremor, as first described by James Parkinson in 1817.1 Although autonomic dysfunction, pain, sleep disturbances, and depression have been recognized as common symptoms in PD their impact on quality of life has been relatively neglected.2

PD cannot be diagnosed until motor symptoms appear but many patients will in hindsight recall a prodromal phase including nonmotor symptoms (NMS).3 Gonera et al. also reported that more nonspecific complaints were reported by PD patients to their general practitioners in the 5 years leading up to diagnosis than were reported by age matched controls.4

The aims of this study were to review the prevalence of NMS as presenting complaints in pathologically-proven cases of PD.



We performed a retrospective review of cases of pathologically-proven PD archived at the Queen Square Brain Bank for Neurological Diseases. Donors were from the UK and died between 1988 and 2005. Of the 543 cases identified 110 were eliminated because of inadequate clinical detail of the presenting features. The London Multi-Centre Research Ethics Committee has approved procedures for the donation of brains to the Queen Square Brain Bank as well as retention of and access to clinical records.

Data Collection

A systematic chart review was performed paying particular attention to the case notes of the family doctor in the 3 years before the documentation of the first motor symptom definitively linked to progressive Parkinsonism. First and subsequent correspondence between the medical specialist and family doctor were also scrutinized. Hospital inpatient notes, inpatient consultations, and emergency room admission notes were also reviewed. The initial presenting complaint prompting patients to attend medical services was noted, and early diagnoses made were documented. The included symptoms or doctors' diagnoses were those considered retrospectively by SOS and DRW to be associated with PD. That is, they did not resolve or they were related to the eventual motor problems, for example; frozen shoulder or sensory disturbances with the later development of bradykinesia, tremor, or rigidity on that same side; depression that did not improve; urinary symptoms with subsequent autonomic dysfunction; new onset low back pain; personality or cognitive changes. We compared patients according to two groups; namely those presenting with exclusively nonmotor and those presenting with motor symptoms. In patients with more than one presenting complaint documented, if motor symptoms were present in addition to NMS as first complaints, these patients were included in the “motor symptoms” group for analysis. The latency from presenting complaint to final clinical diagnosis was reviewed. Other clinical features were recorded as present or absent early in the disease (within 2 years of first symptom onset) and at anytime during the disease. These features included: age of onset (age at the time of the first reported symptom considered to be attributable to disease), disease duration (time from onset until death), falls, bradykinesia, cognitive dysfunction (reported by the family, patient, or physician, and not attributed to affective disorder), symmetry of disease at onset, tremor, rigidity, postural reflexes (reported by the physician), response to levodopa (L-dopa) (graded by patient or clinician from 1 to 4, 1 = nil or slight, 2 = moderate, 3 = good, or 4 = excellent), and autonomic dysfunction. Abnormal autonomic function was recorded according to test results or reports of any two of these symptoms: urinary urgency, frequency, and nocturia without hesitancy; chronic constipation; postural hypotension; sweating abnormalities; erectile dysfunction. The presence and time of onset of hallucinations and dyskinesias were noted. Sleep disturbances were not included in the analysis. We recorded symptoms as absent if not reported and clinical signs as missing if they were not specifically mentioned in the notes. Where reports of clinical features conflicted, the findings of the neurologist were used. Complete drug history was obtained, including latency from symptom onset to initiation, and time from initiation to maximum dose of L-dopa.


Comparison between groups was performed using the Student's t test or Mann-Whitney U test as appropriate. Statistical analysis of difference in the frequency of categorical variables was performed using the χ2 test or Fisher Exact Probability Tests as appropriate. Statistical analyses of data were performed with SPSS version 12.0 (SPSS, Chicago, IL).


From the 433 cases with detailed histories of early symptoms, 91 (21%) had exclusively NMS at presentation to their general practitioner. Of the NMS, pain was the most frequent, seen in 48 (53%) of these cases, urinary symptoms were present in 15 (16.5%), depression or anxiety in 11 (12.1%). “Other” NMS included non specific “cognitive impairment” (5.5%) without functional limitation or dementia, lethargy (4.4%), sensory disturbances, and visual impairment. (Table 1)

Table 1. Patient characteristics and presenting symptoms
 Total cases (N = 433)Cases with “motor symptoms” at presentation (N = 342)Cases without “motor symptoms” at presentation (N = 91)P value
  • a


  • b

    Student's t-test.

  • c

    Mann-Whitney U.

Male:female (%)274:159 (63:37%)215:127 (63:37%)59:32 (65:35%)NSa
Number of patients with a documented family history of PD (%)31 (7%)28 (8.2%)3 (3.3%)0.08a
Age of PD onset mean ± SD60.9 ± 10.4 years60.6 ± 10.6 years61.9 ± 9.5 yearsNSb
Interval between symptom onset and diagnosis of PD; median, interquartile range (years)1.1 (0.9–2.4)1.0 (0.8–2.2)1.6 (1.0–3.0)0.001c
Duration of PD before death; mean ± SD (years)14.9 ± 6.915.3 ± 7.013.4 ± 6.60.016b
Age of death; mean ± SD75.8 ± 7.475.9 ± 7.675.3 ± 6.6NSb
First symptoms including tremor196 (45.3%)196 (57.3%)0N/A
First symptoms including bradykinesia136 (31.4%)136 (39.8%)0N/A
First symptoms including rigidity44 (10.2%)44 (12.9%)0N/A
First symptoms including unspecified gait disturbance51 (11.8%)51 (14.9%)0N/A
First symptoms including pain65 (15%)17 (5%)48 (52.7%)N/A
First symptoms including urinary dysfunction17 (3.9%)2 (0.6%)15 (16.5%)N/A
First symptoms including depression or anxiety11 (2.5%)011 (12.1%)N/A
Other symptoms59 (13.6%)34 (9.9%)25 (27.5%)N/A

Presenting with NMS is associated with a delayed diagnosis of PD, with a median interval of 1.6 years between first symptom and diagnosis of PD, compared with 1.0 years in those presenting with motor symptoms (Mann-Whitney U, P = 0.001). PD was the initial diagnosis made in only 15 (16.5%) of patients presenting with NMS, compared to 230 (67.3%) for patients in the motor group (χ2, P < 0.0001). The nonmotor cases were more likely to be diagnosed with osteoarthritis, degenerative spinal disease, frozen shoulder, depression, or anxiety. (Table 2) Correspondingly, patients presenting with NMS were significantly less likely to be initially referred to a neurologist by their general practitioner than the motor group (5.5% vs. 44.2%, χ2P < 0.0001). Patients in the nonmotor group were more likely to be referred to orthopedic and rheumatological services (29.7%), urological services (15.4%), general physicians (13.2%), and psychiatrists (11%) than a neurologist. In contrast, 26% of patients from the motor group were referred to general physicians and only 4.3% to orthopaedic or rheumatological services. A high proportion of patients presenting with NMS underwent surgical interventions prior to the diagnosis of PD, including steroid injections in 9 of 10 patients diagnosed with frozen shoulders, and five out of 12 patients diagnosed with degenerative spinal disease.

Table 2. Initial diagnoses made following symptom onset
 Total cases (N = 433)Cases with “motor symptoms” at presentation (N = 342)Cases without “motor symptoms” at presentation (N = 91)P value
  • a


PD245 (56.6%)230 (67.3%)15 (16.5%)<0.0001a
Osteoarthritis/degenerative verterbral disease21 (4.8%)9 (2.6%)12 (13.2%)<0.0001a
Depression, anxiety or psychogenic causes17 (3.9%)7 (2%)10 (11%)0.0003a
Frozen shoulder14 (3.2%)3 (0.7%)11 (12.1%)<0.0001a
Stroke9 (2.1%)7 (2%)2 (2.2%)NSa
Essential tremor8 (1.8%)8 (2.3%)0 (0%)NSa
Other diagnoses57 (13.2%)23 (6.7%)34 (37.4%) 
Unknown62 (14.3%)55 (16.1%)7 (7.7%)NSa

A small percentage of patients from the nonmotor group (4.4%) had no bradykinesia, rigidity, or tremor documented 2 years after symptom onset. After the same interval, only two of the 342 patients (0.6%) had none of the “classical parkinsonian triad” documented, both patients having presented with unspecified gait disturbances or writing difficulties (P = 0.02 df = 1, χ2 analysis). See Table 3 for symptoms documented within 2 years from symptom onset. Autonomic dysfunction was the only symptom that was more common in the nonmotor group (P = 0.025, df = 1, χ2 analysis). Tremor was more frequently documented throughout the disease in the motor group (P = 0.04, df = 1, χ2 analysis). There were no other differences in the frequency of clinical features between the groups. Hallucinations were documented in 168 (52.2%) of the motor group and 46 (53.5%) of the nonmotor group during their disease history. There was no difference in the mean latency from first symptom onset to the development of hallucinations between groups (motor group 131 ± 75 months, nonmotor 115 ± 81 months). Drug-induced dyskinesias were documented in 158 (53%) of the motor group, and 41 (50%) in the nonmotor group during the disease history. The mean latency to onset of dyskinesias from first symptom onset was similar in both groups (motor group 101 ± 52 months, nonmotor group 93 ± 45).

Table 3. Symptoms documented at 2 years from first symptom
 Total cases (N = 433)Cases with “motor symptoms” at presentationCases without “motor symptoms” at presentationP value*
  • *


Bradykinesia354 (85.3%)278 (85%)76 (86.4%)NS
Tremor308 (73.9%)251 (76.3%)57 (64.8%)0.04
Extra-axial rigidity321 (78.3%)254 (78.2%)67 (78.8%)NS
Postural instability36 (9%)25 (7.8%)11 (13.9%)NS
Falls23 (5.5%)17 (5.2%)6 (6.7%)NS
Cognitive impairment42 (10%)29 (8.8%)13 (14.6%)NS
Autonomic dysfunction29 (7%)15 (4.6%)14 (16.1%)0.001
Depression97 (24.7%)71 (22.9%)26 (31.3%)NS
Dysarthria50 (12.3%)39 (12.1%)11 (13.4%)NS
Hallucinations12 (5.9%)8 (5.0%)4 (8.9%)NS
Dyskinesia7 (3.6%)6 (3.8%)1 (2.9%)NS

The final clinical diagnosis before death was PD in 81 of 91 (89%) patients initially presenting with NMS, compared with 316 of 342 (92.4%) patients presenting with motor symptoms (P > 0.05). There were no significant differences in the latency to initiation of L-dopa, the maximum doses prescribed, or the response to this medication between these groups. (Table 4) Similarly, no differences were seen between groups regarding the prescription of dopamine agonists. Dopamine agonists were used in 52% of patients presenting with motor symptoms, compared to 46% of the patients presenting with NMS, (P = 0.5). Nonergot agonists were prescribed in ∼80% of these cases in both patient groups. Antidepressant medication prescription rates throughout the disease course were also similar between patient groups, at 29.2% of patients presenting with motor symptoms, compared to 29.7% of the patients presenting with NMS.

Table 4. History of L-dopa use and response
 Total casesCases with “motor symptoms” at presentationCases without “motor symptoms” at presentationP value*
  • *

    Student's t-test.

Latency from first symptom to L-dopa commencement; mean ± SD2.9 ± 2.8 years2.9 ± 2.92.9 ± 2.2NS
Maximum L-dopa dose (mg/day); mean ± SD923 ± 499947 ± 502828 ± 478NS
Response to L-dopa (%)395 (96%)299 (94.9%)79 (98.8%)NS
Grade of response to L-dopa (QSBB scale); mean ± SD3.2 ± 0.93.2 ± 0.93.3 ± 0.9NS


It is thought that by the time patients fulfill diagnostic criteria for the diagnosis of PD, there may be degeneration of ∼30% of neurons in the substantia nigra.5, 6 The duration of the premotor phase following the onset of nigral loss is uncertain, but pathological and radiological estimates have suggested between 5 and 7 years.6, 7

In this study 21% of people who went on to develop motor features of PD, described their initial symptoms in exclusively nonmotor terms, and pain was seen as a presenting symptom in 15%. We acknowledge the retrospective nature of the study is a limiting factor and the recording and interpretation of nonspecific nonmotor symptoms in general practice may vary. General practitioners in the United Kingdom are however required by law to make a written entry of all symptom complaints for every consultation. We accept nonetheless that some physicians may be more aware and astute in suspecting early Parkinsonism when a nonmotor presentation occurs. Hyposmia is rarely spontaneously reported by patients with PD and REM sleep disorder may be dismissed. Once the diagnosis of PD has been firmly established however direct enquiry about deterioration of olfaction frequently results in a positive response We also acknowledge that among the high percentage of pain as a presenting symptom, some could be pain related to bradykinesia in akinesia dominant PD, which would be more likely to be mechanical pain related to a rigid arm rather than the typical pain of PD. Nevertheless, we feel that this study is relevant as it is the first to assess the prevalence and clinical characteristics of a subgroup of patients with pathologically proven PD who present with NMS.

Autonomic dysfunction has been described in up to 50% of patients with diagnosed PD,8 and one study on constipation has been associated with an almost threefold increased risk of a subsequent diagnosis of PD after 12-year follow-up.9 In our cohort, new urinary symptoms were found to be a presenting complaint in 3.9% of patients with PD. Autonomic dysfunction was found to be the only nonmotor symptom that remained more prevalent throughout the course of the PD in the group of patients presenting initially with NMS, compared to those presenting with motor-related complaints. Shorter disease duration is found in patients presenting with NMS (P = 0.0016, t-test). Whether this is due to more frequent and severe autonomic dysfunction in this group is unclear, as in the majority of our cases formal autonomic function tests were not performed, and the specific cause of death is not documented.

Depression or anxiety was shown to be a presenting complaint in 2.5% of our PD cases, with ∼25% developing these symptoms within two years of disease onset. These findings are in keeping with previous research showing that depression affects between 30 and 50% of people with PD10, 11 Studies have suggested that neuropsychiatric conditions may also be “preclinical” symptoms of PD, with strong associations between premorbid depression and subsequent development of PD.12–14 Similarly, anxiety has been associated with subsequent diagnoses of PD,14, 15 although the evidence is not as strong as for depression.13

Hyposmia has been associated with PD for over 30 years, and has been widely studied with clinical and pathological evidence of olfactory dysfunction in a majority of patients.16 The presence of idiopathic hyposmia is thought to predict the development of clinical PD in around 10% of patients at 2-year follow-up.17 Sleep disturbance, including REM sleep behavior disorder (RBD) precede the development of the classical parkinsonian symptoms in up to 38% of cases, with a latency of over 10 years between RBD onset and clinical Parkinsonism.18, 19 Excessive day-time sleepiness has also been shown to precede the diagnosis of PD, with a threefold increased risk demonstrated in one longitudinal study.20 Unfortunately there was almost no documentation of the presence or absence of olfactory or sleep disturbances in patients we studied, possibly due to a lack of awareness of the clinical importance of these features during the period in which this cohort was treated.

Nonmotor presentations of PD were frequently misdiagnosed initially by primary carers, leading to a high proportion of potentially inappropriate specialist referrals and treatments, including steroid injections for frozen shoulders and surgical interventions for degenerative spinal disease or carpal tunnel syndromes. Clearly, such interventions may have a deleterious effect on patients' quality of life, in addition to potentially increasing costs of health-care. We identified an increased latency between symptom onset and final diagnosis of PD in these patients. However, despite these early diagnostic difficulties, we have shown that, after the first 2 years from symptom onset, these groups have similar symptoms and signs, and the rate of correct final clinical diagnoses did not differ. A good response to L-dopa therapy was seen in both subgroups, despite suggestions that NMS tend to respond less well than motor symptoms to dopaminergic treatments.21 However, we acknowledge that the response rating to L-dopa is an overall clinical impression, and does not account for individual symptoms that may not be deemed significant by the treating clinicians. Furthermore the limitations of the retrospective nature our data, and the selection bias that is expected in a brain bank postmortem series, may account for the differences with previous studies.22 The large number of patients and the inclusion of patients that otherwise might not have been included in a prospective clinical study are relative strengths of this study.

Despite the increasing literature demonstrating that NMS are frequent and disabling features of more advanced PD,23 they are still likely to be underrecognized by neurologists.24 Our findings suggest that NMS may be significant features in earlier PD stages and that an increased awareness of these problems as manifestations of PD is also required at a primary care level. To minimize misdiagnoses and potentially harmful and inappropriate interventions in patients presenting without the “typical” motor symptoms, a clearer message about NMS of early PD in medical education is required. This may be facilitated by the development of clinical tools such as the validated Non Motor Questionnaire to aid clinicians in identifying, and managing, nonmotor symptoms in patients at earlier stages of PD.25


This study was supported by the Reta Lila Weston Trust for Medical Research. Without the kind donations from the patients registered with the Queen Square Brain Bank for Neurological Disorders, this study would not have been possible.