Parkinsonism

Parkinson disease remains a clinical diagnosis. No accurate laboratory or radiologic test is currently available. An acute challenge with levodopa, followed by monitoring for improvement of symptoms, is not useful diagnostically.

Prior Probability

Parkinson disease affects 1% of people older than 65 years and 2% of those older than 85 years.² The prevalence among older patients presenting with general neurologic complaints is almost certainly much higher, but precise figures are unavailable.

Population for Whom Parkinson Disease Should Be Considered

Adults with a tremor or other symptoms noted in Tables 38-4 and 38-5 (see Figure 38-2 for assessment of bradykinesia).

Detecting the Likelihood of Parkinson Disease

See Tables 38-4 and 38-5 (see Figure 38-2 for assessment of bradykinesia).

Reference Standard Tests

Serial clinical examinations performed by a specialist.

Clinical Scenario

A 68-year-old man presents with a 3-month history of right arm tremor at rest. His movements have been slower and he has difficulty getting out of a chair. Physical examination reveals rigidity in the upper limbs. He walks with small steps and has limited ability to swing his arms. His facial expressions are limited.

Why Answer This Question With a Clinical Examination?

With a prevalence estimated between 150 and 200 per 100 000, Parkinson disease (PD) is one of the most common neurologic disorders.¹ It is more prevalent in older persons, affecting 1% of those older than 65 years and 2% of those older than 85 years.²

Although common, the diagnosis of PD is challenging. Laboratory tests are not available and conventional imaging studies are not helpful. The best reference standard is, unfortunately, neuropathologic (depletion of brain stem pigmented neurons and proliferation of Lewy bodies).³ Serial neurologic evaluation with or without concomitant treatment can also be used.⁴ The response to an acute levodopa challenge has been used as a diagnostic tool. This test is problematic for a number of reasons: its sensitivity and specificity are low, acute levodopa administration is associated with significant adverse effects, there is lack of agreement on what constitutes a threshold response, and the test is expensive and inconvenient.⁵ The clinical examination, therefore, is the basis for initial diagnosis. Classic clinical features of PD include tremor at rest, bradykinesia, rigidity, and postural instability.

There is evidence that the accuracy of diagnosis in some settings is improving. In a 1991 study by Rajput et al⁶ among 41 patients diagnosed clinically with PD by neurologists, the disease was confirmed neuropathologically at autopsy in 31 (positive predictive value [PPV] of 76%). Hughes et al⁷ evaluated the accuracy of clinical diagnosis among 100 patients with PD, 86 of whom were followed up by neurologists, 7 by geriatricians, and 7 by internists. The diagnosis was confirmed at autopsy in 90 persons (PPV = 90%). Another study confirmed PD at autopsy among 72 of 73 patients (PPV = 99%) followed up by neurologists affiliated with a highly specialized movement disorders center.⁸ Despite these improvements and impressive results, it is important to keep in mind that the clinical diagnoses in these studies were often made during a long period and by physicians with a great deal of expertise and experience. The accuracy of clinical diagnosis in other settings is unclear. PD is still mistaken for other neurologic disorders. The most frequent misdiagnoses include progressive supranuclear palsy, multisystem atrophy (MSA) (encompasses the diagnoses Shy Drager syndrome, olivopontocerebellar atrophy, and striatonigral degeneration), and dementia with Lewy bodies.⁶ The differential diagnosis also includes essential tremor and vascular pseudoparkinsonism. Mistaking PD for other conditions can lead to inappropriate and ineffective treatment. Although a patient with essential tremor, for example, may benefit from a β-blocker, this treatment would have no effect on the tremor of PD. Inappropriate treatment based on misdiagnosis also delays the use of dopaminergic medications, which can decrease the severity of symptoms and disability.⁹

Mistaking other disorders for PD is also harmful. Dyskinesias, for example, appear in 15% to 85% of persons within 5 years of treatment with levodopa, and hallucinations occur in 20% of patients.¹⁰ There is also evidence that levodopa causes damage to dopamine neurons, leading to accelerated dopamine degeneration.⁵ Whether the initial diagnosis is correct or not, the disease has serious social and psychological consequences.^{11, 12} In summary, the clinical examination is important in suspected PD because no laboratory or radiologic tests are helpful diagnostically. Misdiagnosis of PD is associated with adverse effects.

Pathophysiologic Characteristics

It is important to distinguish between PD and parkinsonism. Parkinsonism refers to any clinical syndrome in which 2 or more features are present such as tremor, rigidity, and bradykinesia. Parkinson disease is a form of primary or idiopathic parkinsonism. Viral infections, environmental toxins, oxidative stress, and heredity have all been suspected as causes.¹³ Secondary or acquired parkinsonism has a variety of causes, including head trauma, cerebrovascular disease, and hydrocephalus.^{14, 15} Secondary parkinsonism may persist for months after the drugs that caused it are discontinued. A thorough inquiry into past and current medication use, therefore, is essential when questioning patients presenting with parkinsonism. Parkinson disease begins as neurons and dopamine are lost from the substantia nigra and intracytoplasmic inclusions (Lewy bodies) appear. Symptoms appear when 70% to 80% of dopamine is lost.¹⁶

Symptoms and Signs

Nonspecific insidious symptoms, including generalized malaise, easy fatigability, and subtle personality changes, mark the onset of PD. These may occur years before the appearance of tremor, limb rigidity, bradykinesia, and postural instability.¹⁶ Numerous secondary manifestations appear unpredictably and are as varied as disordered sleep (42% of patients),¹⁷ constipation (50%), pain (50%), depression (40%), and dementia (20%).^{16, 18} Signs typically begin unilaterally and then progress asymmetrically.

James Parkinson described the combination of tremor and bradykinesia as a shaking palsy.¹⁸ Seventy-five percent of patients complain initially of a tremor that usually occurs at rest in an upper extremity and is characterized by visible oscillations with a frequency of 4 to 6 per second. Tremor appears intermittently, disappearing during sleep and increasing in severity during times of emotional distress or anxiety. It is often described as pill-rolling, because a rhythmic movement is observed in the hand as the index finger flexes and extends against the thumb repetitively.¹⁹

Some basic features distinguish the tremor of PD from physiologic and essential tremors (Boxes 38-1 and 38-2).^{20, 21}

Rigidity, refers to an involuntary stiffness of the skeletal muscles and is a common sign. Electromyogram assessment of parkinsonian patients reveals an alternating discharge pattern in opposing muscle groups, even at rest (eg, triceps and biceps). Resistance to movement of limbs may be smooth or interrupted. Cog wheeling refers to the jerky motion of limbs as constant force is applied across a joint, which is similar to the ratcheting of the cogs of gears as they click.²² Unlike rigidity, spasticity refers to a selective increase of tone of flexor muscles in the arms and extensor muscles in the legs²³ and suggests a diagnosis other than PD.

Bradykinesia refers to the overall slowing of active movement or slowness in initiating movement. The initial surge of motor activity is inadequate and movements are fragmented into a series of incremental steps. Postural instability in patients with PD presents as changes in gait and balance. Short and shuffling steps are often accompanied by festination. Loss of arm movements commonly appears. The patient may walk with the arms straight down, rather than swinging them back and forth. Gait disturbance is the major cause of disability in many patients. As postural reflex mechanisms are lost, patients become stooped and have a tendency to fall. Those with severe deficits are sometimes confined to a wheelchair or bed.

How to Elicit Signs

Tremor

Tremor can be defined as any rhythmic, involuntary oscillatory movement of a body part. The tremor classification is complex and has overlapping features in different disease syndromes. Nevertheless, the Movement Disorder Society has developed a classification system to help clinicians distinguish tremor types.²⁴ Tremors can be classified as rest or action.

The classification system divides tremors into 11 syndromes. Patients with PD typically have a slow (frequency of about 4 6/s) tremor at rest. It is easily observed by having patients position their hands on their lap. Physicians should be able to identify the key features of PD and essential and physiologic tremors (Box 38-2).

Precise measurement of tremor frequency and amplitude is sometimes used in diagnostic evaluation. This requires special devices and is beyond the scope of the clinical examination.

Rigidity

Involuntary muscle stiffness or rigidity can be shown if resistance to passive movement of the limbs is detected. With the patient relaxed, the examiner places his or her thumb across the antecubital fossa with one hand while passively flexing and extending the elbow several times with the other hand. Rigidity often increases with repeated flexion and extension movements. With cog wheeling, the examiner feels alternate periods of resistance and relaxation. With lead pipe rigidity, the examiner feels smooth but increased muscle tone throughout passive flexion and extension.²⁵ Rigidity and cog wheeling may be felt in other large joints, but if detected in the arms, there is no need to confirm their presence elsewhere. Many patients with essential tremor manifest a rhythmic resistance to passive movements of a limb while there is voluntary action of another body part. This is not true cog wheeling but is known as Froment sign, which also appears in PD patients.²⁶

Bradykinesia

Bradykinesia refers to a decrease in the speed and amplitude of complex movements. Jobbagy et al²⁷ described 4 maneuvers designed to detect bradykinesia: tapping the fingers, twiddling, pinching and circling, and tapping with the heel (Figure 38-1). Twiddlin grefers to repeated rotation of the hands in front of the body. The pinching and circling test is a sequence of 6 movements: pinching (opposing thumb and index finger) with the right hand and then with the left hand; circling (rotating the hand in a circle) with the right hand and then with the left hand; pinching with the right hand while simultaneously circling with the left; and pinching with the left hand while simultaneously circling with the right (Figure 38-1). Jobbagy et al²⁷ were able to quantify the performance of patients on these tasks by using a motion analyzer, although a specific threshold “score” to define bradykinesia was not determined. However, poor performance of these maneuvers is easily detectable and clinicians can use them to confirm the presence of bradykinesia subjectively.

Glabella Tap Reflex

This reflex is tested by percussion of the forehead with the examiner's index finger or by pulling a fold of skin between the thumb and index finger on the temple lateral to the external canthus and tapping with the thumb. The orbicularis oculi muscle reflexively contracts, causing both eyes to blink. The reflex blinking normally stops after tapping is repeated 5 to 10 times. Persistent blinking is a positive response sometimes referred to as Myerson sign.²⁸ Care should be taken to keep the examiner's finger above the patient's eyes to avoid blinking in response to visual threat (Figure 38-2).

Are These Features Found in Other Diseases?

The symptoms and signs of idiopathic PD overlap with those of other neurologic diseases, including MSA and progressive supranuclear palsy.

Like PD, MSA often presents with asymmetric rigidity and akinesia, but only a minority of patients have a resting tremor.²⁹ Half of patients with MSA present with autonomic dysfunction and cerebellar symptoms, and one-quarter demonstrate a transient response to levodopa.^{25, 29} Similarly, patients with progressive supranuclear palsy seldom present with tremor. Rigidity and postural instability, however, are common.³⁰

Parkinsonism is sometimes also a feature of Alzheimer disease.¹⁴ However, Alzheimer disease is easy to distinguish from PD because other features are much more prominent. Furthermore, unlike in PD, cognitive impairment is present at the onset of Alzheimer disease.

Methods

Four of the authors (G.R., L.F., T.O., and C.E.) performed independent searches of the MEDLINE database (1966 2001), using a number of Medical Subject Headings (“exp tremor,” “exp PD,” “essential tremor”) combined with the search terms and strategy used for The Rational Clinical Examination series.³¹

All relevant articles were retrieved. The resulting set of articles was divided into 3 parts, each of which was reviewed by a pair of authors. The reference lists of all articles were also carefully searched for additional articles. Articles were included for study if they met the following criteria: dealt primarily with the diagnosis of PD; included patients presenting with 1 or more typical parkinsonian symptoms or signs (eg, tremor, rigidity); final diagnosis confirmed by a suitable criterion standard, such as serial or detailed neurologic evaluation or pathologic confirmation at autopsy; and contained original data from which 2 × 2 tables could be extracted to calculate the sensitivity, specificity, positive likelihood ratio (LR+), and negative likelihood ratio (LR–) for different signs and symptoms. Because the number of suitable articles was small, additional inclusion criteria such as a minimum sample size or publication after a certain year were not used. However, the quality of articles included was assessed according to criteria previously developed for this series.³¹

The likelihood ratios (LRs) for different diagnostic features were calculated when not available in the original articles. Corresponding 95% confidence intervals (CIs) were determined by the method of Greenland and Robins.³² All values were rounded to 2 significant digits. When identical or similar diagnostic features appeared in more than 1 article and the patients were similar across studies in terms of demographics and illness characteristics, weighted summary LRs (pooled LRs) and the corresponding 95% CIs were calculated with the DerSimonian Laird random-effects method.³³ We used MetaWin statistical software (version 2; Sinauer Associates, Sunderland, Massachusetts).

Results

Quality of the Evidence

A total of 185 articles were reviewed. All authors agreed about which articles met our selection criteria. We chose 6 articles. Two articles^{34, 35} included independent blind comparisons of symptoms and signs of a small number of patients who had been diagnosed as having PD or MSA according to comparison of clinical records to pathologic results at autopsy. Because the patients studied had already been diagnosed clinically as having PD or MSA, selection bias was a serious problem. These 2 articles provided level 3 evidence, leading to grade C recommendations (see Table 1-7 for summary of Evidence Grades and levels).

The remaining 4 articles^{36, 37, 38, and 39} had numerous methodologic biases. Although of lower methodologic quality, they can still be classified as containing level 3 evidence and providing grade C recommendations (Table 38-1). Selection bias was a major problem in all 4 articles because many of the patients evaluated had either been diagnosed as having PD on initial clinical examination or had obvious parkinsonian features. In one study,³⁹ a screening instrument was administered to patients with an initial diagnosis of PD, peripheral neuropathy, stroke, or epilepsy. The instrument included both a self administered questionnaire and a set of physical tasks, performance of which was graded subjectively. Neurologists confirming the presence of PD were aware of each patient's initial diagnosis and responses to the screening instrument. This obviously makes the study prone to observer bias. Like most studies with low methodologic quality, these 4 articles^{36, 37, 38, and 39} reported optimistic LRs.

Precision

Interclinician and intraclinician reliability of symptoms and signs was documented only for the glabella tap sign.³⁶ Precision could not be quantified in the clinicopathologic studies because symptom histories were obtained retrospectively from medical records.

Interclinician reliability in eliciting the glabella tap sign was found to be 88% among patients with intracranial disease and 100% in controls.³⁶ A κ coefficient for interclinician agreement could not be calculated because data about how each clinician scored each patient were not included. No causes for imprecision in assessing symptoms or signs were documented in the selected articles.

Accuracy

Several symptoms, collected by patient self-report in a questionnaire,³⁸ significantly increase the likelihood of PD when present and decrease it when absent. The symptoms are trouble turning in bed, shuffling while walking, micrographia, difficulty rising from a chair, loss of balance, and trouble opening jars. The diagnostic value of tremor as a symptom varied widely among the selected articles, with a range in LRs+ of 1.3 to 11 (Table 38-2).

The lack of tremor as a symptom makes PD less likely (range of LRs–, 0.24-0.60). However, the usefulness of the lack of tremor as a symptom is limited by verification bias in the corresponding studies. Verification bias occurs when confirmatory or criterion standards are selectively applied to patients, depending on the results of their preliminary screening test.⁴⁰ The independent value of tremor detected on neurological examination has an LR+ of only 1.5 (95% CI, 1.0-2.3), while the absence of a tremor detected on examination about halves the likelihood of PD (LR–, 0.47; 95% CI, 0.27-0.84) (Table 38-3).³⁴

Rigidity as a symptom has an LR+ range of 1.3 to 4.5 and makes PD more likely. The absence of rigidity has a broad LR, making it less useful (LR– range, 0.12-0.73) (Table 38-2). As a sign detected on neurological examination (Table 38-3), rigidity as an independent value is more useful (LR+, 2.8; 95% CI, 1.8-4.4; LR–, 0.38; 95% CI, 0.19-0.76).³⁹ When both rigidity and bradykinesia are present, the LR+ for the combination of findings improves to 4.5 (95% CI, 2.9-7.1), while the absence of both findings makes PD much less likely, with an LR– of 0.12 (95% CI, 0.03-0.45).³⁹

The glabella tap sign is useful, with an LR+ of 4.5 (95% CI, 2.8 7.4) and an LR– of 0.13 (95% CI, 0.03 0.47).³⁶ Changes in voice have an LR range of 2.6 to 3.7, while the lack of a voice change makes PD somewhat less likely (LR– range, 0.25-0.73).^{37, 39} The results confirm the limited usefulness of the response to levodopa because the LR+ is only 1.2 (95% CI, 0.87-1.6), while the absence of a response has an LR– of 0.63 (95 % CI, 0.31-1.3).³⁴

Physicians sometimes must consider whether patients have PD or multiple systems atrophy (MSA), so a series of findings have been compared between the 2 disorders.³⁵ Patients with rigidity as an initial presenting feature of PD are less likely to have PD (LR, 0.53; 95% CI, 0.35-0.80) and more likely to have MSA. The presence of dementia also favors PD over MSA (LR+, 3.2; 95% CI, 1.5-6.8). Not surprisingly, central or autonomic nervous systems findings are much less likely with PD (LR range when central or autonomic findings present, 0.03-0.31), so their presence favors MSA. Bradykinesia and symptoms of depression do not help distinguish between the disease (95% CI for the LR includes 1).

The Bottom Line

Few studies address the clinical diagnosis of PD rigorously. Nearly 200 years after it was first described, the accurate clinical diagnosis of PD remains a significant challenge. There is a great need for diagnostic studies involving larger numbers of patients in which presenting symptoms and signs are prospectively compared with the final diagnosis, established through a suitable criterion standard, such as autopsy or serial neurologic evaluation.

A number of classic features of PD, when present, do help establish the diagnosis. These include the symptoms of tremor, the combination of rigidity and bradykinesia, loss of balance, micrographia, and shuffling gait. Difficulty with the tasks of turning in bed, opening jars, and rising from a chair should also raise the suspicion of PD. It is difficult to gauge the usefulness of the absence of tremor as a symptom in ruling out PD because of verification bias in the studies in which it was evaluated.

The diagnostic value of the classic combination of tremor, rigidity, and bradykinesia on examination is modest at best. Useful signs include the glabella tap, difficulty walking heel to toe, and the presence of rigidity on examination.

Author Affiliations at the Time of the Original Publication

Department of Family Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania (Drs Rao, Fisch, and Srinivasan); University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania (Drs D’Amico, Okada, and Eaton); and Kaiser Permanente of Colorado, Aurora (Dr Robbins).

Acknowledgments

We thank Richard Bedlack, MD, Mitchell Heflin, MD, and Amy Rosenthal, MD, for their helpful reviews of the manuscript; Amy Haugh, MLS, for her valuable library assistance; and David L. Simel, MD, MHS, for his valuable guidance.

Clinical Scenario—Resolution

This patient presents with many common features of PD. You can question him about the tasks of turning in bed and opening jars. A sample of his writing may reveal micrographia. A glabella tap test should be performed. Additional positive symptoms or signs would justify empiric treatment with dopaminergic medication, with careful follow-up by a physician experienced in the treatment of this condition.

References

Clinical Scenario

A 70-year-old man presents with a 3-month history of progressive tremor and difficulty in writing letters. His wife states that he has trouble eating with a spoon. Physical examination reveals tremor in both hands. Otherwise, he feels well and continues to play golf.

Updated Summary on Parkinson Disease

Original Review

Updated Literature Search

Our literature search used the parent search strategy for The Rational Clinical Examination series, combined with the subject headings “exp tremor,” “exp Parkinson disease,” and “essential tremor,” published in English from 2002 to July 2004. No new original articles address the sensitivity and specificity of clinical findings for Parkinson disease (PD). We found 1 new systematic review that addresses key questions related to the diagnosis and management of PD. Much of the recent research in the diagnosis of PD has focused on functional neuroimaging with positron emission tomography and single-photon-emission computed tomography.

Changes in the Reference Standard

There are no laboratory or radiographic reference standards for PD. The best reference standard is serial clinical assessments performed by a specialist. Newer imaging methods may eventually be routinely helpful in identifying PD at an early stage or in patients with atypical signs and symptoms and in following the progression of disease. However, they have not yet been evaluated with enough methodologic rigor to allow clinicians to be confident in the results.

New Findings

One systematic review confirms that PD remains a clinical diagnosis and that neuroimaging, the levodopa challenge, and other diagnostic tests are not useful.¹

Improvements in the Data Presented in the Original Publication

Tables 38-2 and 38-3 were revised because they originally included data for helping to distinguish PD from multiple systems atrophy. These tables now show only the likelihood ratios for making the diagnosis of PD.

Results of Literature Review

No data suggest new useful findings for PD. Symptoms common in PD include tremor, rigidity, bradykinesia, and micrographia (unusually small handwriting). Difficulty with tasks such as turning in bed, opening jars, or rising from a chair are also common. The most useful clinical findings are the glabella tap (see Figure 38-1) and heel-to-toe tests.

Evidence From Guidelines

No guidelines explicitly address the diagnosis of PD.

Clinical Scenario—Resolution

The patient should be asked about key symptoms of PD, including loss of balance (positive likelihood ratio [LR+], 1.6-6.6), shuffling gait (LR+, 3.3-15), and rigidity or stiffness (LR+, 1.3-4.5). It is not clear why he has trouble eating with a spoon. This could be a result of his tremor or because of slowness of movements (bradykinesia). The patient should be asked about difficulty rising from a chair (LR+, 1.9-5.2), a specific manifestation of bradykinesia. His difficulty in writing letters could also be attributed to tremor, bradykinesia, or micrographia.

Further evaluation reveals that the patient's tremor is of slow frequency (4-6/s) and occurs at rest, which is typical of the tremor of PD. By contrast, classic essential tremor is usually postural (occurs while maintaining a position against gravity) or kinetic (occurs during voluntary movements.) Physiologic tremor is usually postural and of higher frequency (8-12/s) than the tremor of PD.

The glabella test should be performed (LR+, 4.5), and whether the patient is able to walk heel to toe (LR+, 2.9) should also be assessed. Asking the patient to rise from a chair and checking to see whether his upper limbs are rigid during passive movement is also useful in establishing the diagnosis (combination of bradykinesia and rigidity; LR+, 4.5).

Parkinson disease is a progressive disease, and patients present with different combinations of clinical features of various severity at different stages. It is therefore, important to evaluate the patient periodically before establishing the diagnosis and initiating treatment.

References for the Update

Clinical Scenario

A 70-year-old man presents with a 3-month history of progressive tremor and difficulty in writing letters. His wife states that he has trouble eating with a spoon. Physical examination reveals tremor in both hands. Otherwise, he feels well and continues to play golf.

Clinical Scenario—Resolution

The patient should be asked about key symptoms of PD, including loss of balance (positive likelihood ratio [LR+], 1.6-6.6), shuffling gait (LR+, 3.3-15), and rigidity or stiffness (LR+, 1.3-4.5). It is not clear why he has trouble eating with a spoon. This could be a result of his tremor or because of slowness of movements (bradykinesia). The patient should be asked about difficulty rising from a chair (LR+, 1.9-5.2), a specific manifestation of bradykinesia. His difficulty in writing letters could also be attributed to tremor, bradykinesia, or micrographia.

Further evaluation reveals that the patient's tremor is of slow frequency (4-6/s) and occurs at rest, which is typical of the tremor of PD. By contrast, classic essential tremor is usually postural (occurs while maintaining a position against gravity) or kinetic (occurs during voluntary movements.) Physiologic tremor is usually postural and of higher frequency (8-12/s) than the tremor of PD.

The glabella test should be performed (LR+, 4.5), and whether the patient is able to walk heel to toe (LR+, 2.9) should also be assessed. Asking the patient to rise from a chair and checking to see whether his upper limbs are rigid during passive movement is also useful in establishing the diagnosis (combination of bradykinesia and rigidity; LR+, 4.5).

Parkinson disease is a progressive disease, and patients present with different combinations of clinical features of various severity at different stages. It is therefore, important to evaluate the patient periodically before establishing the diagnosis and initiating treatment.