Parkinsonism and Related Disorders 21 (2015) 477e482
Contents lists available at ScienceDirect
Parkinsonism and Related Disorders journal homepage: www.elsevier.com/locate/parkreldis
Cardiovascular autonomic testing performed with a new integrated instrumental approach is useful in differentiating MSA-P from PD at an early stage Francesca Baschieri a, 1, Giovanna Calandra-Buonaura a, b, 1, Andrea Doria a, Francesca Mastrolilli c, Aldopaolo Palareti d, Giorgio Barletta a, b, Laura Solieri a, b, Pietro Guaraldi e, Paolo Martinelli a, Pietro Cortelli a, b, * a
Department of Biomedical and Neuromotor Sciences (DIBINEM), University of Bologna, Bologna, Italy IRCCS, Institute of Neurological Sciences, Bellaria Hospital, Bologna, Italy c Department of Neurology, Lewisham and Greenwich NHS Trust, Queen Elizabeth Hospital, London, UK d Department of Computer Science and Engineering, University of Bologna, Bologna, Italy e Public Local Health Authority of Modena and Bologna, Italy b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 September 2014 Received in revised form 29 January 2015 Accepted 13 February 2015
Objective: To evaluate whether a battery of cardiovascular autonomic tests (Ewing's battery, EB) performed with a new integrated instrumental approach is useful in differentiating multiple system atrophy with predominant parkinsonism (MSA-P) from Parkinson's disease (PD) at an early stage. Methods: We retrospectively analyzed EB tests of all the patients (n ¼ 99) with a parkinsonian syndrome referred to our clinic who performed EB during the first diagnostic workup and were subsequently evaluated at least once a year until a final diagnosis of MSA-P (n ¼ 34) or PD (n ¼ 65). Thirty-eight controls matched for age and sex were included. EB consisted of head-up tilt test (HUTT), Valsalva manoeuvre (VM), deep breathing, and sustained handgrip whose correct execution and results were checked and obtained automatically. Results were compared between groups. Discriminant analysis was performed to identify MSA-P or PD patients. Results: Orthostatic hypotension was found in 22 MSA-P and 3 PD patients. Cardiovascular reflexes indices were significantly more affected in MSA-P compared to PD and controls. EB presented a 91% sensitivity and 94% specificity in the differentiation of MSA-P and PD. HUTT þ VM presented a 91% sensitivity and 92% specificity. Conclusions: Our results suggest that EB or HUTT þ VM performed with an integrated instrumental approach and analyzed with the discriminant procedure may distinguish MSA-P from PD at an early stage and might improve the accuracy of current diagnostic criteria. However, a validation in separate samples and prospective studies is needed. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Multiple system atrophy Parkinson's disease Disorders of autonomic nervous system Cardiovascular autonomic testing Orthostatic hypotension
1. Introduction
* Corresponding author. IRCCS, Institute of Neurological Sciences, Department of Biomedical and Neuromotor Sciences, University of Bologna, c/o Padiglione G, Ospedale Bellaria, via Altura n 3, 40139 Bologna, Italy. Tel.: þ39 051 4966929. E-mail addresses: [email protected] (F. Baschieri), giovanna. [email protected] (G. Calandra-Buonaura), [email protected] (A. Doria), [email protected] (F. Mastrolilli), [email protected] (A. Palareti), [email protected] (G. Barletta), [email protected] (L. Solieri), [email protected] (P. Guaraldi), [email protected] (P. Martinelli), [email protected] (P. Cortelli). 1 These authors contributed equally to the manuscript. http://dx.doi.org/10.1016/j.parkreldis.2015.02.011 1353-8020/© 2015 Elsevier Ltd. All rights reserved.
In clinical practice, it is of crucial importance to distinguish multiple system atrophy with predominant parkinsonism (MSA-P) from Parkinson's disease (PD) at disease onset, due to the different treatment and prognosis they require [1,2]. Impairment of the autonomic control of the cardiovascular system with a predominant failure of the sympathetic branch of the autonomic nervous system (autonomic failure, AF) has been observed in these diseases, affecting mainly the preganglionic component in MSA and the postganglionic one in PD [3].
478
F. Baschieri et al. / Parkinsonism and Related Disorders 21 (2015) 477e482
Orthostatic hypotension (OH), defined as a sustained reduction of systolic blood pressure (SBP) of at least 20 mm Hg or diastolic blood pressure (DBP) of at least 10 mm Hg within 3 min of standing or head-up tilt test (HUTT) to at least 60 on a tilt table [4], is the cardinal sign of AF and may be symptomatic or asymptomatic. Diagnosis of OH caused by AF requires exclusion of non-neurogenic conditions and confirmation through at least cardiovascular reflex tests [5]. According to consensus criteria, cardiovascular AF, defined only by the presence of OH, or urinary AF is mandatory for the diagnosis of MSA [6]. In MSA, OH whose estimated prevalence is 75% [1] may occur early in the disease course or late as in long-surviving patients [7,8]. In PD a recent meta-analysis reported a 30% prevalence at any stage of the disease with a large statistical heterogeneity between studies [9]. Although early and severe autonomic involvement is an exclusion criterion of PD [10], autopsy-proven PD patients, presenting OH before the appearance of motor signs, have been described [11] and the sole presence of OH did not allow the distinction between MSA and PD [12e16]. However, previous studies assessing whether cardiovascular AF evaluated by means of cardiovascular reflex tests could help in the differential diagnosis of these diseases gave contrasting results [12e14,17e20]. We aimed to evaluate whether a battery of non-invasive, inexpensive and reproducible cardiovascular autonomic tests (Ewing's battery [21], EB) performed with a new integrated instrumental approach [22] and analysed with an advanced statistical procedure (discriminant analysis) is useful in differentiating MSA-P from PD at disease onset when diagnosis is still uncertain even in a movement disorders clinic. 2. Materials and methods 2.1. Patients and control subjects We retrospectively evaluated the cardiovascular autonomic tests of all the patients fulfilling the following criteria: 1) referral to the Movement Disorders clinic of the University of Bologna between 2000 and 2013 for a non-vascular parkinsonian syndrome of uncertain diagnosis at disease onset; 2) absence of metabolic and cardiovascular diseases and drugs that could affect autonomic testing; 3) mini mental status examination >24; 4) written informed consent to personal data processing for research purposes; 5) correct execution of the EB with an integrated instrumental method [22] during the first diagnostic workup independently from orthostatic symptoms; 6) subsequent evaluation at least once a year until a final clinical diagnosis of MSA-P or PD. Three neurologists expert in movement disorders, blinded to EB results, independently confirmed the diagnosis according to consensus criteria [6,10] by reviewing all data collected from the first to the last available follow-up visit performed in our department. EB results were retrospectively abstracted from medical records by another author who in turn was blinded to the final clinical diagnosis. Thirty-four patients eventually diagnosed with MSA-P and 65 patients with PD were included in the study. We also included 38 age- and sex-matched controls with negative family history of parkinsonian syndromes, negative personal history of cardiovascular, metabolic, endocrine, renal diseases or other conditions that can possibly interfere with autonomic testing, normal neurological examination and EB results. The study was approved by the DIBINEM institutional review board. 2.2. Cardiovascular reflexes Tests were performed in the morning in a temperature-controlled clinical investigation room (23 ± 1 C). Patients had been drug free overnight and all subjects were allowed to have only a light breakfast avoiding coffee and tea. The following parameters were acquired continuously during the tests by means of a polygraph amplifier (Model 15LT, Grass Techonologies, Quincy, MA): heart rate (HR), oronasal and abdominal breathing and peripheral vasomotor tone with finger photoelectric transducer. Beat to beat SBP and DBP were monitored continuously (Finometer Midi, Finapres Medical Systems, Amsterdam, The Netherlands). All signals were collected by means of a specific analogic to digital converter and a piece of software (SparkBio Srl, Bologna, Italy) able to visualize, store, and analyse the data, providing a final report with the results. After 30 min of supine rest, HUTT (10 min at 65 ), Valsalva manoeuvre (VM; 40 mm Hg for 15 s), deep breathing (DB; 6 breaths/ min), and sustained handgrip (HG; 1/3 of maximal effort for 5 min) were performed. The correct execution of each test was checked automatically by an electronic device and a specialized technician who monitored the subjects' level of vigilance. The
manoeuvres were carried out under video-monitoring, allowing a period of rest required to reach basal BP and HR values in-between investigations. This integrated instrumental method of autonomic evaluation automatically calculated the following parameters: 1) basal SBP, DBP and HR as the mean value of the last 5 min of supine rest preceding HUTT; 2) response to HUTT as the difference (D) between values at 3 min and basal values; 3) Valsalva ratio (VR) ¼ HR in phase II/ HR in phase IV; 4) presence of BP recovery in late phase II of the VM (D BP II_L) ¼ mean BP in late phase II (II_L) e mean BP in early phase II (II_E); 5) presence of overshoot in phase IV of the VM (D BP IV) ¼ mean BP phase IV e mean BP in phase I; 6) sinus arrhythmia during DB (D IE) ¼ average of the 10 shortest ReR intervals during inspiration e average of the 10 longest ReR during expiration; 7) response to HG as D with respect to the basal values of SBP, DBP and HR after 5 min of isometric effort [22]. During the same assessment, patients were asked about symptoms of orthostatic intolerance during daily life (i.e. postural dizziness or fainting on standing erect or after meals or exercise, visual disturbances, neck ache radiating to the occipital region of the skull and to the shoulders, recovery of all symptoms after lying flat) while data on urinary symptoms and sudomotor function were not collected systematically. 2.3. Statistical analysis Data were analysed using the Statistical Analysis Software (SAS; version 9.3, Cary, NC, USA). GLM procedure was used for comparison of clinical characteristics and EB results among groups. Influence of age, sex, body mass index (BMI) and type of disease on basal values was also tested. Statistical significance was set at p 0.05. Linear discriminant analysis was performed to verify whether EB might identify PD and MSA-P patients. The function used for the discriminant analysis was calculated on the basis of a parametric method based on a multivariate normal distribution within each class that uses the pooled covariance matrix in calculating the squared distances for testing the hypothesis that the class means are equal in the population for each variable. The discriminant analysis generates a linear function whose result gives the probability of being classified as PD (not MSA-P) if this probability is greater than 50%. ROC curves and area under the curve statistics will be presented. EB results of 34 MSA-P and 65 PD patients were used as training set. EB parameters considered for the analysis were: 1) supine rest SBP, DBP and HR; 2) HUTT: D SBP, DBP and HR; 3) VM: VR, D BP II_L, and D BP IV; 4) DB: D IE; 5) HG: D SBP, DBP and HR. Discriminant analysis was performed considering all the variables for each test. Tests were analysed both separately and in combination (EB). Basal values were considered in every case.
3. Results 3.1. Population Between 2000 and 2013, 843 patients were referred for the first time to our Movement Disorders clinic, 391 of whom for a parkinsonian syndrome. 227 (58%) patients were excluded because of concurrent diseases (e.g. diabetes) and/or medications that interfered with autonomic testing, 28 (7%) patients did not give the consent and 17 (4%) were lost at follow up. 119 (30%) patients were finally considered, but 20 did not correctly execute all the tests of EB and were not included in the analysis because their results were not reliable. Such patients were similar for age and sex with the remaining group of 99 patients and at follow-up 5 of them turned out to be MSA. Clinical characteristics of patients considered in the study are shown in Table 1. Data are expressed as mean and standard deviation. At EB there were no significant differences among groups in age (MSA-P and PD ¼ 61 ± 10 and controls ¼ 55 ± 10 years) and BMI. Disease duration from motor onset to EB was similar between PD and MSA-P patients (2.9 ± 2.3 and 3.2 ± 2.0 years), while the subsequent follow-up was longer for PD patients (5.8 ± 4.2 and 3.5 ± 2.3 years). Hoehn and Yahr stage was 1 or 2 in 62/65 PD and in 20/34 MSA-P patients. Only a minority of patients was under treatment with levodopa or dopamine agonists. In treated patients there was no significant difference in levodopa equivalent daily dose (LED) [23]. Eighteen patients with MSA-P (53%) and 11 patients with PD (17%) reported symptoms suggestive of OH. None had already been treated for OH.
F. Baschieri et al. / Parkinsonism and Related Disorders 21 (2015) 477e482 Table 1 Patient characteristics at autonomic testing.
Males BMI (Kg/m2)a Age at investigation (y)a Age at motor onset (y)a Disease duration (y)a Follow up (y)a Hoehn & Yahr Stage 1 Stage 2 Stage 3 Symptoms of OH Levodopa Pramipexole Ropinirole Rotigotine Untreated LED (mg)b
479
Table 3 Discriminant analysis results.
MSA-P (n ¼ 34)
PD (n ¼ 65)
24 (71) 26.3 ± 4.3 (20.3e35.3) 61 ± 10 (40e80) 58 ± 10 (33e76) 3.2 ± 2.0 (1e8) 3.5 ± 2.3 (1e9)
40 (62) 26.3 ± 4.1 (19.6e37.3) 61 ± 11 (38e85) 58 ± 11 (36e81) 2.9 ± 2.3 (1e8) 5.8 ± 4.2 (1e13)
5 (15) 15 (44) 14 (41) 18 (53) 15 (44) 2 (6) e 1 (3) 16 (47) 325 (100e635)
37 (57) 25 (38) 3 (5) 11 (17) 25 (38) 9 (14) 2 (3) e 29 (45) 375 (100e810)
Figures are number (%) unless stated otherwise. MSA-P, multiple system atrophy with predominant parkinsonism; PD, Parkinson's disease; BMI, body mass index; n, number of patients; y, years; disease duration, from motor onset to investigation with EB; OH, orthostatic hypotension; LED, levodopa equivalent daily dose. a Data are expressed as mean ± sd (range). b Data are expressed as median (range).
3.2. Autonomic testing We did not find any significant influence of age, sex and BMI on supine rest DBP and HR values; supine SBP was influenced only by age (p < 0.002). Basal values and EB results were significantly different between MSA-P and controls (Table 2). At supine rest, only DBP was significantly different between PD and controls while all EB parameters were significantly different with the exception of HR response to HUTT and HG. Baseline values were significantly different in MSA-P and PD, with the exception of DBP. We found OH [4] in 20/34 MSA-P and in 3/65 PD patients. At VM, DB and HG, all cardiovascular responses were significantly more affected in MSA-P (p < 0.0001).
HUTT VM DB HG EB HUTT þ VM
Sensitivity (CI)
Specificity (CI)
79% 91% 88% 85% 91% 91%
91% 92% 83% 94% 94% 92%
(0.62e0.91) (0.76e0.98) (0.73e0.97) (0.69e0.95) (0.76e0.98) (0.76e0.98)
(0.81e0.97) (0.83e0.97) (0.72e0.91) (0.85e0.98) (0.83e0.97) (0.85e0.98)
HUTT, head-up tilt test; VM, Valsalva manoeuvre; DB, deep breathing; HG, sustained handgrip; EB, Ewing battery; CI, 95% confidence interval (Exact or Clopper-Pearson confidence limits).
Among patients who complained of orthostatic symptoms, OH was confirmed in 15/18 MSA-P and 2/11 PD patients. Five MSA-P and 1 PD patients had asymptomatic OH. 3.3. Discriminant analysis The discriminant function for the complete battery of cardiovascular autonomic tests is: F13.5642 þ 0.101892*SBP supine rest e 0.14681*DBP supine rest þ 0.165726*HR supine rest þ 0.023932*D SBP HUTT e 0.15861*D DBP HUTT þ 0.031092*D HR HUTT þ 2.962883*VR e 0.03009*D BP IV e 0.09869*D BP II_L e 0.15121*D IE DB e 0.0892*D SBP HG e 0.05564*D DBP HG e 0.0526*D HR HG. The cut-off is 0.0 (P ¼ 0.50), negative values are PD and positive values are MSA. Considering each test, HUTT had the least discriminant power between MSA-P and PD (Table 3), whereas the best discriminating test was the VM which correctly identified 31/34 MSA-P and 60/65 PD patients. The combination of HUTT þ VM performed similarly to the VM alone. The discriminant function for HUTT þ VM is: F ¼ 13.1617 þ 0.08538*SBP supine rest e 0.12559*DBP supine rest þ 0.16466*HR supine rest e 0.00242*D SBP HUTT e 0.13119*D DBP HUTT e 0.01332*D HR HUTT þ 0.73742*VR e 0.0623*D BP IV e 0.07302*D BP II_L. With the complete battery only one more PD patient was correctly classified (61/65) (Fig. 1A). Fig. 1B and C shows the receiver operating characteristics (ROC) curve for the whole battery and for the combination of HUTT þ VM. Discriminant analysis performed for patients without symptoms of OH correctly identified 13/16 (81%) MSA-P and 53/54 (98%) PD patients.
Table 2 Autonomic test results.
HUTT Supine rest SBP (mm Hg) Supine rest DBP (mm Hg) Supine rest HR (bpm) D SBP (mm Hg) D DBP (mm Hg) D HR (bpm) OH consensus n Valsalva VR D BP II_L (mm Hg) D BP IV (mm Hg) Deep breathing D IE (bpm) Handgrip D SBP (mm Hg) D DBP (mm Hg) D HR (bpm)
MSA-P (n ¼ 34)
PD (n ¼ 65)
Controls (n ¼ 38)
p-value MSA-P vs PD
p-value MSA-P vs controls
p-value PD vs controls
141 ± 19 76 ± 15 74 ± 11 26 ± 24 10 ± 14 10 ± 9 20 (59%)
123 ± 19 75 ± 12 65 ± 10 3 ± 13 7±8 13 ± 8 3 (5%)
119 ± 10 64 ± 8 64 ± 7 11 ± 11 11 ± 6 12 ± 6 0
Contents lists available at ScienceDirect
Parkinsonism and Related Disorders journal homepage: www.elsevier.com/locate/parkreldis
Cardiovascular autonomic testing performed with a new integrated instrumental approach is useful in differentiating MSA-P from PD at an early stage Francesca Baschieri a, 1, Giovanna Calandra-Buonaura a, b, 1, Andrea Doria a, Francesca Mastrolilli c, Aldopaolo Palareti d, Giorgio Barletta a, b, Laura Solieri a, b, Pietro Guaraldi e, Paolo Martinelli a, Pietro Cortelli a, b, * a
Department of Biomedical and Neuromotor Sciences (DIBINEM), University of Bologna, Bologna, Italy IRCCS, Institute of Neurological Sciences, Bellaria Hospital, Bologna, Italy c Department of Neurology, Lewisham and Greenwich NHS Trust, Queen Elizabeth Hospital, London, UK d Department of Computer Science and Engineering, University of Bologna, Bologna, Italy e Public Local Health Authority of Modena and Bologna, Italy b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 September 2014 Received in revised form 29 January 2015 Accepted 13 February 2015
Objective: To evaluate whether a battery of cardiovascular autonomic tests (Ewing's battery, EB) performed with a new integrated instrumental approach is useful in differentiating multiple system atrophy with predominant parkinsonism (MSA-P) from Parkinson's disease (PD) at an early stage. Methods: We retrospectively analyzed EB tests of all the patients (n ¼ 99) with a parkinsonian syndrome referred to our clinic who performed EB during the first diagnostic workup and were subsequently evaluated at least once a year until a final diagnosis of MSA-P (n ¼ 34) or PD (n ¼ 65). Thirty-eight controls matched for age and sex were included. EB consisted of head-up tilt test (HUTT), Valsalva manoeuvre (VM), deep breathing, and sustained handgrip whose correct execution and results were checked and obtained automatically. Results were compared between groups. Discriminant analysis was performed to identify MSA-P or PD patients. Results: Orthostatic hypotension was found in 22 MSA-P and 3 PD patients. Cardiovascular reflexes indices were significantly more affected in MSA-P compared to PD and controls. EB presented a 91% sensitivity and 94% specificity in the differentiation of MSA-P and PD. HUTT þ VM presented a 91% sensitivity and 92% specificity. Conclusions: Our results suggest that EB or HUTT þ VM performed with an integrated instrumental approach and analyzed with the discriminant procedure may distinguish MSA-P from PD at an early stage and might improve the accuracy of current diagnostic criteria. However, a validation in separate samples and prospective studies is needed. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Multiple system atrophy Parkinson's disease Disorders of autonomic nervous system Cardiovascular autonomic testing Orthostatic hypotension
1. Introduction
* Corresponding author. IRCCS, Institute of Neurological Sciences, Department of Biomedical and Neuromotor Sciences, University of Bologna, c/o Padiglione G, Ospedale Bellaria, via Altura n 3, 40139 Bologna, Italy. Tel.: þ39 051 4966929. E-mail addresses: [email protected] (F. Baschieri), giovanna. [email protected] (G. Calandra-Buonaura), [email protected] (A. Doria), [email protected] (F. Mastrolilli), [email protected] (A. Palareti), [email protected] (G. Barletta), [email protected] (L. Solieri), [email protected] (P. Guaraldi), [email protected] (P. Martinelli), [email protected] (P. Cortelli). 1 These authors contributed equally to the manuscript. http://dx.doi.org/10.1016/j.parkreldis.2015.02.011 1353-8020/© 2015 Elsevier Ltd. All rights reserved.
In clinical practice, it is of crucial importance to distinguish multiple system atrophy with predominant parkinsonism (MSA-P) from Parkinson's disease (PD) at disease onset, due to the different treatment and prognosis they require [1,2]. Impairment of the autonomic control of the cardiovascular system with a predominant failure of the sympathetic branch of the autonomic nervous system (autonomic failure, AF) has been observed in these diseases, affecting mainly the preganglionic component in MSA and the postganglionic one in PD [3].
478
F. Baschieri et al. / Parkinsonism and Related Disorders 21 (2015) 477e482
Orthostatic hypotension (OH), defined as a sustained reduction of systolic blood pressure (SBP) of at least 20 mm Hg or diastolic blood pressure (DBP) of at least 10 mm Hg within 3 min of standing or head-up tilt test (HUTT) to at least 60 on a tilt table [4], is the cardinal sign of AF and may be symptomatic or asymptomatic. Diagnosis of OH caused by AF requires exclusion of non-neurogenic conditions and confirmation through at least cardiovascular reflex tests [5]. According to consensus criteria, cardiovascular AF, defined only by the presence of OH, or urinary AF is mandatory for the diagnosis of MSA [6]. In MSA, OH whose estimated prevalence is 75% [1] may occur early in the disease course or late as in long-surviving patients [7,8]. In PD a recent meta-analysis reported a 30% prevalence at any stage of the disease with a large statistical heterogeneity between studies [9]. Although early and severe autonomic involvement is an exclusion criterion of PD [10], autopsy-proven PD patients, presenting OH before the appearance of motor signs, have been described [11] and the sole presence of OH did not allow the distinction between MSA and PD [12e16]. However, previous studies assessing whether cardiovascular AF evaluated by means of cardiovascular reflex tests could help in the differential diagnosis of these diseases gave contrasting results [12e14,17e20]. We aimed to evaluate whether a battery of non-invasive, inexpensive and reproducible cardiovascular autonomic tests (Ewing's battery [21], EB) performed with a new integrated instrumental approach [22] and analysed with an advanced statistical procedure (discriminant analysis) is useful in differentiating MSA-P from PD at disease onset when diagnosis is still uncertain even in a movement disorders clinic. 2. Materials and methods 2.1. Patients and control subjects We retrospectively evaluated the cardiovascular autonomic tests of all the patients fulfilling the following criteria: 1) referral to the Movement Disorders clinic of the University of Bologna between 2000 and 2013 for a non-vascular parkinsonian syndrome of uncertain diagnosis at disease onset; 2) absence of metabolic and cardiovascular diseases and drugs that could affect autonomic testing; 3) mini mental status examination >24; 4) written informed consent to personal data processing for research purposes; 5) correct execution of the EB with an integrated instrumental method [22] during the first diagnostic workup independently from orthostatic symptoms; 6) subsequent evaluation at least once a year until a final clinical diagnosis of MSA-P or PD. Three neurologists expert in movement disorders, blinded to EB results, independently confirmed the diagnosis according to consensus criteria [6,10] by reviewing all data collected from the first to the last available follow-up visit performed in our department. EB results were retrospectively abstracted from medical records by another author who in turn was blinded to the final clinical diagnosis. Thirty-four patients eventually diagnosed with MSA-P and 65 patients with PD were included in the study. We also included 38 age- and sex-matched controls with negative family history of parkinsonian syndromes, negative personal history of cardiovascular, metabolic, endocrine, renal diseases or other conditions that can possibly interfere with autonomic testing, normal neurological examination and EB results. The study was approved by the DIBINEM institutional review board. 2.2. Cardiovascular reflexes Tests were performed in the morning in a temperature-controlled clinical investigation room (23 ± 1 C). Patients had been drug free overnight and all subjects were allowed to have only a light breakfast avoiding coffee and tea. The following parameters were acquired continuously during the tests by means of a polygraph amplifier (Model 15LT, Grass Techonologies, Quincy, MA): heart rate (HR), oronasal and abdominal breathing and peripheral vasomotor tone with finger photoelectric transducer. Beat to beat SBP and DBP were monitored continuously (Finometer Midi, Finapres Medical Systems, Amsterdam, The Netherlands). All signals were collected by means of a specific analogic to digital converter and a piece of software (SparkBio Srl, Bologna, Italy) able to visualize, store, and analyse the data, providing a final report with the results. After 30 min of supine rest, HUTT (10 min at 65 ), Valsalva manoeuvre (VM; 40 mm Hg for 15 s), deep breathing (DB; 6 breaths/ min), and sustained handgrip (HG; 1/3 of maximal effort for 5 min) were performed. The correct execution of each test was checked automatically by an electronic device and a specialized technician who monitored the subjects' level of vigilance. The
manoeuvres were carried out under video-monitoring, allowing a period of rest required to reach basal BP and HR values in-between investigations. This integrated instrumental method of autonomic evaluation automatically calculated the following parameters: 1) basal SBP, DBP and HR as the mean value of the last 5 min of supine rest preceding HUTT; 2) response to HUTT as the difference (D) between values at 3 min and basal values; 3) Valsalva ratio (VR) ¼ HR in phase II/ HR in phase IV; 4) presence of BP recovery in late phase II of the VM (D BP II_L) ¼ mean BP in late phase II (II_L) e mean BP in early phase II (II_E); 5) presence of overshoot in phase IV of the VM (D BP IV) ¼ mean BP phase IV e mean BP in phase I; 6) sinus arrhythmia during DB (D IE) ¼ average of the 10 shortest ReR intervals during inspiration e average of the 10 longest ReR during expiration; 7) response to HG as D with respect to the basal values of SBP, DBP and HR after 5 min of isometric effort [22]. During the same assessment, patients were asked about symptoms of orthostatic intolerance during daily life (i.e. postural dizziness or fainting on standing erect or after meals or exercise, visual disturbances, neck ache radiating to the occipital region of the skull and to the shoulders, recovery of all symptoms after lying flat) while data on urinary symptoms and sudomotor function were not collected systematically. 2.3. Statistical analysis Data were analysed using the Statistical Analysis Software (SAS; version 9.3, Cary, NC, USA). GLM procedure was used for comparison of clinical characteristics and EB results among groups. Influence of age, sex, body mass index (BMI) and type of disease on basal values was also tested. Statistical significance was set at p 0.05. Linear discriminant analysis was performed to verify whether EB might identify PD and MSA-P patients. The function used for the discriminant analysis was calculated on the basis of a parametric method based on a multivariate normal distribution within each class that uses the pooled covariance matrix in calculating the squared distances for testing the hypothesis that the class means are equal in the population for each variable. The discriminant analysis generates a linear function whose result gives the probability of being classified as PD (not MSA-P) if this probability is greater than 50%. ROC curves and area under the curve statistics will be presented. EB results of 34 MSA-P and 65 PD patients were used as training set. EB parameters considered for the analysis were: 1) supine rest SBP, DBP and HR; 2) HUTT: D SBP, DBP and HR; 3) VM: VR, D BP II_L, and D BP IV; 4) DB: D IE; 5) HG: D SBP, DBP and HR. Discriminant analysis was performed considering all the variables for each test. Tests were analysed both separately and in combination (EB). Basal values were considered in every case.
3. Results 3.1. Population Between 2000 and 2013, 843 patients were referred for the first time to our Movement Disorders clinic, 391 of whom for a parkinsonian syndrome. 227 (58%) patients were excluded because of concurrent diseases (e.g. diabetes) and/or medications that interfered with autonomic testing, 28 (7%) patients did not give the consent and 17 (4%) were lost at follow up. 119 (30%) patients were finally considered, but 20 did not correctly execute all the tests of EB and were not included in the analysis because their results were not reliable. Such patients were similar for age and sex with the remaining group of 99 patients and at follow-up 5 of them turned out to be MSA. Clinical characteristics of patients considered in the study are shown in Table 1. Data are expressed as mean and standard deviation. At EB there were no significant differences among groups in age (MSA-P and PD ¼ 61 ± 10 and controls ¼ 55 ± 10 years) and BMI. Disease duration from motor onset to EB was similar between PD and MSA-P patients (2.9 ± 2.3 and 3.2 ± 2.0 years), while the subsequent follow-up was longer for PD patients (5.8 ± 4.2 and 3.5 ± 2.3 years). Hoehn and Yahr stage was 1 or 2 in 62/65 PD and in 20/34 MSA-P patients. Only a minority of patients was under treatment with levodopa or dopamine agonists. In treated patients there was no significant difference in levodopa equivalent daily dose (LED) [23]. Eighteen patients with MSA-P (53%) and 11 patients with PD (17%) reported symptoms suggestive of OH. None had already been treated for OH.
F. Baschieri et al. / Parkinsonism and Related Disorders 21 (2015) 477e482 Table 1 Patient characteristics at autonomic testing.
Males BMI (Kg/m2)a Age at investigation (y)a Age at motor onset (y)a Disease duration (y)a Follow up (y)a Hoehn & Yahr Stage 1 Stage 2 Stage 3 Symptoms of OH Levodopa Pramipexole Ropinirole Rotigotine Untreated LED (mg)b
479
Table 3 Discriminant analysis results.
MSA-P (n ¼ 34)
PD (n ¼ 65)
24 (71) 26.3 ± 4.3 (20.3e35.3) 61 ± 10 (40e80) 58 ± 10 (33e76) 3.2 ± 2.0 (1e8) 3.5 ± 2.3 (1e9)
40 (62) 26.3 ± 4.1 (19.6e37.3) 61 ± 11 (38e85) 58 ± 11 (36e81) 2.9 ± 2.3 (1e8) 5.8 ± 4.2 (1e13)
5 (15) 15 (44) 14 (41) 18 (53) 15 (44) 2 (6) e 1 (3) 16 (47) 325 (100e635)
37 (57) 25 (38) 3 (5) 11 (17) 25 (38) 9 (14) 2 (3) e 29 (45) 375 (100e810)
Figures are number (%) unless stated otherwise. MSA-P, multiple system atrophy with predominant parkinsonism; PD, Parkinson's disease; BMI, body mass index; n, number of patients; y, years; disease duration, from motor onset to investigation with EB; OH, orthostatic hypotension; LED, levodopa equivalent daily dose. a Data are expressed as mean ± sd (range). b Data are expressed as median (range).
3.2. Autonomic testing We did not find any significant influence of age, sex and BMI on supine rest DBP and HR values; supine SBP was influenced only by age (p < 0.002). Basal values and EB results were significantly different between MSA-P and controls (Table 2). At supine rest, only DBP was significantly different between PD and controls while all EB parameters were significantly different with the exception of HR response to HUTT and HG. Baseline values were significantly different in MSA-P and PD, with the exception of DBP. We found OH [4] in 20/34 MSA-P and in 3/65 PD patients. At VM, DB and HG, all cardiovascular responses were significantly more affected in MSA-P (p < 0.0001).
HUTT VM DB HG EB HUTT þ VM
Sensitivity (CI)
Specificity (CI)
79% 91% 88% 85% 91% 91%
91% 92% 83% 94% 94% 92%
(0.62e0.91) (0.76e0.98) (0.73e0.97) (0.69e0.95) (0.76e0.98) (0.76e0.98)
(0.81e0.97) (0.83e0.97) (0.72e0.91) (0.85e0.98) (0.83e0.97) (0.85e0.98)
HUTT, head-up tilt test; VM, Valsalva manoeuvre; DB, deep breathing; HG, sustained handgrip; EB, Ewing battery; CI, 95% confidence interval (Exact or Clopper-Pearson confidence limits).
Among patients who complained of orthostatic symptoms, OH was confirmed in 15/18 MSA-P and 2/11 PD patients. Five MSA-P and 1 PD patients had asymptomatic OH. 3.3. Discriminant analysis The discriminant function for the complete battery of cardiovascular autonomic tests is: F13.5642 þ 0.101892*SBP supine rest e 0.14681*DBP supine rest þ 0.165726*HR supine rest þ 0.023932*D SBP HUTT e 0.15861*D DBP HUTT þ 0.031092*D HR HUTT þ 2.962883*VR e 0.03009*D BP IV e 0.09869*D BP II_L e 0.15121*D IE DB e 0.0892*D SBP HG e 0.05564*D DBP HG e 0.0526*D HR HG. The cut-off is 0.0 (P ¼ 0.50), negative values are PD and positive values are MSA. Considering each test, HUTT had the least discriminant power between MSA-P and PD (Table 3), whereas the best discriminating test was the VM which correctly identified 31/34 MSA-P and 60/65 PD patients. The combination of HUTT þ VM performed similarly to the VM alone. The discriminant function for HUTT þ VM is: F ¼ 13.1617 þ 0.08538*SBP supine rest e 0.12559*DBP supine rest þ 0.16466*HR supine rest e 0.00242*D SBP HUTT e 0.13119*D DBP HUTT e 0.01332*D HR HUTT þ 0.73742*VR e 0.0623*D BP IV e 0.07302*D BP II_L. With the complete battery only one more PD patient was correctly classified (61/65) (Fig. 1A). Fig. 1B and C shows the receiver operating characteristics (ROC) curve for the whole battery and for the combination of HUTT þ VM. Discriminant analysis performed for patients without symptoms of OH correctly identified 13/16 (81%) MSA-P and 53/54 (98%) PD patients.
Table 2 Autonomic test results.
HUTT Supine rest SBP (mm Hg) Supine rest DBP (mm Hg) Supine rest HR (bpm) D SBP (mm Hg) D DBP (mm Hg) D HR (bpm) OH consensus n Valsalva VR D BP II_L (mm Hg) D BP IV (mm Hg) Deep breathing D IE (bpm) Handgrip D SBP (mm Hg) D DBP (mm Hg) D HR (bpm)
MSA-P (n ¼ 34)
PD (n ¼ 65)
Controls (n ¼ 38)
p-value MSA-P vs PD
p-value MSA-P vs controls
p-value PD vs controls
141 ± 19 76 ± 15 74 ± 11 26 ± 24 10 ± 14 10 ± 9 20 (59%)
123 ± 19 75 ± 12 65 ± 10 3 ± 13 7±8 13 ± 8 3 (5%)
119 ± 10 64 ± 8 64 ± 7 11 ± 11 11 ± 6 12 ± 6 0
