Parkinsonism

Introduction

Strang's |1~ survey, methodology unspecified, of 897 married sufferers from parkinsonism did not suggest that their spouses were at any increased risk of developing the condition. However, we have been impressed by the number of spouses of elderly sufferers who appear to have parkinsonian features. A diagnosis of parkinsonism requires the presence of two of the four cardinal signs, hypo/bradykinesia, muscle rigidity, a characteristic tremor and postural abnormality |2~. We have |3, 4~ used nocturnal axial rotation whilst recumbent as a measure of the tendency towards parkinsonism with respect to hypo/bradykinesia. Angular displacement was significantly less in elderly sufferers from idiopathic |2~ parkinsonism than in their spouses, and in these spouses than in controls. The angular displacement of the spouses was independent of whether they shared a bed with the sufferer, and neither these nor control couples showed any relationship between partners in angular displacement. Selective mating for poor turning or learned behaviour is unlikely to explain the reduced movement of spouses of sufferers. We have now quantified the four cardinal signs and associated cognitive disability in sufferers, their spouses and control couples.

Methods

Subjects: Twenty sufferers (P) from idiopathic parkinsonism (five men and five women in the age groups 70-79 and 80-89 years) and their spouses (Ps), and 20 control couples, comprising index partners (C), without parkinsonism, and their spouses (Cs) were studied. Ps and Cs could have parkinsonism. P and C were matched for age band and sex. Subjects gave informed consent to participate in the study, which had local Ethics Committee approval.

Overall inclusion and exclusion criteria, applied to consecutively presenting couples, were as follows: all subjects were Caucasian, had English as their first language, were free from cardiovascular and respiratory symptoms during normal daily activities and did not use a walking aid. Those with overt abnormality of lower limbs, a history of orthopaedic surgery to, or with any pain in joints of spine or lower limbs were excluded. A history of specific neurological disorder (other than parkinsonism), musculoskeletal disorder, or a condition which might mimic a cardinal sign, led to exclusion, as did clinical dementia, depression or other mental illness. A cognitive function score |5~ of less than 50% was regarded as compatible with clinical dementia. Subjects receiving hypnotic or sedative drugs were excluded.

Assessments: The clinical assessment included Webster ratings of rigidity and tremor |6~ and measurement of the ranges of ankle plantar flexion and subtalar inversion (see A prognostic index), of cognitive function on a 16-point |5~ and of affect on a 20-point scale |7~.

Distance/time measurements of gait were obtained using the gait-assessment trolley |8, 9~, a computerized method, based on infra-red telemetry, which allows free walking in a non-laboratory environment. Rested subjects walked for up to 40 m or for 1 min in a 2.5 m wide empty corridor, 'at your own speed', following the command 'go'. After 10 min, the walk was repeated, using the shoestring device |10~, which also measures foot separation while walking. Standing body sway was measured as total angular movement in the sagittal plane |11~ during 3 consecutive minutes. Reaction times were measured as the time taken to lift an index finger from its touch-sensitive support, with and without a warning as to whether the instruction would be to lift left or right finger |12~. The delay between the 'Ready' and the 'Go' instruction on the computer screen was random (between 1 and 3 s).

Recordings were processed by a 'blind' observer. Clinical ratings and goniometry were repeated in Ps and controls on a separate occasion by two observers blind to identity and category of subject: mean values for the two observers were analysed.

Statistical methods: Dependent variables in the statistical analysis were classified as (A) descriptive, (B-E) measures of the four cardinal signs of parkinsonism and (F) associated cognitive disability. A one-way analysis of variance was used to investigate the hypotheses that there were no differences between groups P, Ps, C and Cs in the mean values for each dependent variable of sets A to E. Orthogonal contrasts were used to make two comparisons: first to ensure that there were no significant differences between C and Cs, and, second, to compare Ps with (C + Cs). A split-plot analysis of variance was performed on F, mean reaction time. For this variable, subject group and contrasts were between-subject effects, whilst test conditions were within-subject effects. Assumptions of normality and equality of residual variance were checked using the Shapiro-Francia W test |13~ and Bartlett's test |14~, respectively. A |log.sub.e~ transformation was employed to ensure the validity of these assumptions for the hopelessness rating, double support time, sway and reaction time.

For the non-continuous variables, the Kruskal-Wallis test |15~ was used to detect whether there were any differences between subject group medians. Where the initial comparison was of more than two groups, Dunn's procedure |15~ was then used to compare Ps with (C + Cs): the latter procedure tends to give conservative results.

Analysis of covariance was used to investigate whether any differences found between Ps and (C + Cs) with respect to the dependent variables were still present, after adjustment for covariates found to be relevant in other groups of controls |10, 16-18~. These 'predetermined' covariates had been selected from common physical and psychometric variables and measurements relating to spine and lower and measurements relating to spine and lower limbs.

A prognostic index: To be able to predict the probability of being parkinsonian, and its interpretation as a measure of the tendency towards parkinsonism, may be an important tool in investigating the contribution of genetic and environmental influences to the pathogenesis. This approach was developed with respect to one cardinal sign: variables relating to hypo/bradykinesia of gait have been measured in 104 volunteers with |16, 17~ and 144 without |19~ clinical idiopathic parkinsonism, in whom other major influences on gait had been excluded. Distance/time measures of gait and measures of the functional anatomy of spine and lower limbs were subjected to logistic regression modelling. Initial models identified seven important predictors of the presence or absence of clinical parkinsonism. These were adjusted to remove associations with age, height and sex, and a final multiple logistic regression model, the 'prognostic index', was constructed, in which only three (adjusted) variables remained important. These were stride length, msl (mm), and ranges of ankle plantar flexion, apf (|degrees~), and subtalar inversion sti (|degrees~).

Values for the above three variables are adjusted to remove their associations with age (y), height (cm) and sex (male = 0, female = 1):

|msl.sub.adj~ = msl + 2.618(age - 60) - 10.941 (height - 175) |apf.sub.adj~ = apf + 0.202(age - 60) - 3.576 (sex) |sti.sub.adj~ = sti + 0.069(age - 60)

The prognostic index, Ip, is obtained by substituting adjusted values into the equation:

Ip = 16.915 - (0.011 |msl.sub.adj~) - (0.092 |apf.sub.adj~) - (0.176 |sti.sub.adj~)

The predicted probability of being parkinsonian is then obtained by:

exp(Ip)/|1 + exp(Ip)~

The performance of this model as a discriminant function was estimated by calculating the cross-validated, predicted probabilities for the combined group of 248 subjects. It was able to classify subjects as having or not having clinical parkinsonism with a reasonable degree of certainty. In the present study, the probability of being parkinsonian in Ps and controls, and the significance of the difference between groups in the prognostic index, were calculated.

Results

Comparison of index controls, C, and their spouses, Cs, with respect to dependent variable sets A-F, showed no significant differences, except for walking speed, measured by the shoestring device (p = 0.03), and that only before adjustment for covariates: data from C and Cs were, thus, pooled when comparisons with Ps were made. The spouses of sufferers, Ps, and controls, were matched for age and sex, and did not differ significantly in height (p = 0.2) or weight (p = 0.3). Raw initial data for variable sets B-F and the derived prognostic index are shown in Figure 2, for P, Ps, C and Cs. Comparison between Ps and controls showed significant differences with respect to hypo/bradykinesia, postural abnormality and pooled reaction time data. Findings relating to hypo/bradykinesia using the shoestring device reflected those using the trolley, for stride length, walking speed and double support time (p = 0.0001, 0.0004 and 0.006, respectively). The findings for reaction time were similar, irrespective of whether dominant or non-dominant hand was used, and of the presence or absence of a warning.

When blinding was applied to tests prone to observer bias (clinical goniometry and ratings) there were significant differences, Ps versus controls, at the 0.05 level in goniometry |mean difference (SE) -4.27 (2.12) |degrees~ and -2.40 (1.22) |degrees~, for ankle plantar flexion and subtalar inversion, respectively~, the difference in the derived prognostic index remained highly significant (p |is less than~ 0.0001), and a significant (p = 0.01) difference became apparent in rigidity rating. It is uncertain whether the rigidity rating became discriminating because of blinding per se, or because activation of muscle tone in the absence of resting rigidity was taken into account. Webster, himself, described the latter as an excellent way to discover early disease |6~.

After adjustment of objective measurements relating to hypo/bradykinesia and postural abnormality for the predetermined covariates, significant differences between Ps and controls remained. However, adjustment reduced the significance of the slower reaction TABULAR DATA OMITTED time of Ps to a trend. Our previous work |10, 16-18~ has shown cognitive function and depression to be relevant covariates for reaction time, but not for hypo/bradykinesia (except for cognitive function in the case of stride length) or postural abnormality. In this study, could the differences between Ps and controls relating to hypo/bradykinesia and postural abnormality be accounted for by cognitive function |20~ or depression |21~? Although mental test scores tended to be lower and hopelessness ratings higher in Ps, differences were not significant. When mental test score or hopelessness rating was incorporated with the predetermined covariates in the analysis, only the former proved to be significant, and that, again, in the case of stride length alone (p = 0.002 and p = 0.008, for trolley and shoestring methods, respectively). Moreover, the difference in stride length between Ps and controls remained significant after adjustment for all identified covariates.

The P/Ps couples had lived together for a median of 51.5 years (interquartile range 47-55 years), age being a determinant of time together (p = 0.001). After adjustment for the covariate age, there was no relationship, in Ps, between time cohabiting and the dependent variables B-F: this may simply reflect that extremely high cohabitation times were the rule.

Discussion

Marked, multifarious and relevant differences between spouses of parkinsonians and control subjects are reported, which are difficult to account for by selective mating, learned or reactive behaviour. Why has a tendency towards parkinsonism in spouses of sufferers not been described previously? Our first proposition is that work on the pathogenesis has been limited by relying on the gross clinical definition of the syndrome. Taking only frank cases may be to study factors governing severity rather than causation. We, therefore, suggest the need to step back and look at the true distribution of the condition or, even better, that of a pre-clinical state. Our second proposition is that work has been limited by lack of studies in the old, where the condition, and presumably the pre-clinical state, is more common. Most P/Ps couples had been cohabiting for about half a century.

Calne and Langston |22~ suggested that idiopathic parkinsonism is caused by an environmental insult producing a depletion of striatal dopamine, which is superimposed on the loss with age. Barbeau and co-workers |23~ proposed that genetic susceptibility was also important. It is possible that lack of concordance in twin studies |24~ represents differences in exposure to an insult between proband and the co-twin, who was destined to develop a pre-clinical state, or if he/she survived long enough, the syndrome itself. Several possible risk and protective factors |24~ have been suggested, but none of these (apart, perhaps, from stress |25~) seems sufficiently widespread to account for the distribution.

The present work demonstrates the feasibility of defining a functional pre-clinical state with respect to hypo/bradykinesia. The definition now needs to be tightened up using objective measures of the other cardinal signs. Longitudinal follow-up is required to confirm whether spouses of sufferers have a progressive pre-clinical state. Screening of families for microsigns of parkinsonism should resolve whether it is influences in early life, adult life or both, which are of critical importance. Case-control studies could focus down on causative factors.

Acknowledgements

Our thanks go to Mrs Caroline Dore, Head of the Section of Medical Statistics, CRC, for her advice, Mrs Janet Gilbert for her administrative work, Mr Ken and Mrs Stella Johnston for their practical help and Roche Products Ltd, UK for their support.

References

1. Strang RR. Conjugal parkinsonism. Dis Nerv Syst 1967;28:814-15.

2. Quinn NP, Hussain FA. Parkinson's disease. Br Med J 1986;293:379-81.

3. Weller C, Bowes SG, Kirk CAA, et al. Measurement of axial rotation: its relevance to screening for night-time hypokinesia in old age and parkinsonism. Age Ageing 1991;20:3-7.

4. Weller C, Nicholson PW, Dobbs SM, Bowes SG, Purkiss A, Dobbs RJ. Reduced axial rotation in the spouses of sufferers from clinical idiopathic parkinsonism. Age Ageing 1992;21:189-94.

5. Denham MJ, Jefferys P. Modified Tooting Bec Questionnaire. In: Source book of geriatric assessment. vol. 2. Hampshire: Karger, 1984;52.

6. Webster DD. Clinical analysis of the disability of Parkinson's disease. Modern treatments 1968;5:257-82.

7. Beck AT, Beamesderfer A. Assessment of depression: the depression inventory. In: Pichot T, ed. Psychological measurements in psychopharmacology, 7, Modern problems in pharmacopsychiatry. Basel: Karger, 1974; 151-9.

8. Weller C, Dobbs RJ, Tate MA, et al. Monitoring gait characteristics in orthopaedic and neurological patients by infrared telemetry. In: Amlaner CJ, ed. Biotelemetry X. Proc Tenth Int Symp Biotelemetry. Fayetteville: University of Arkansas Press, 1989;575-80.

9. Bowes SG, Clark PK, Charlett A, et al. Objective outcome criteria in trials of anti-parkinsonian therapy in the elderly: reliability, sensitivity and specificity of measures of brady- or hypokinesia. Br J Clin Pharmacol 1991;31:295-304.

10. Weller C, Humphrey SJE, Kirollos C, et al. Gait on a shoestring: falls and foot separation in parkinsonism. Age Ageing 1992;21:242-4.

11. Wright BM. A simple mechanical ataxia-meter. J Physiol (Lond) 1971;218:27-8P.

12. Bowes SG, Frith C, Walker C, et al. Screening for drug effects on reaction time in parkinsonism: methodology. Eur J Clin Pharmacol 1989;36:139.

13. Royston JP. A simple method for evaluating the Shapiro-Francia W test for non-normality. Statistician 1983;32:297-300.

14. Armitage P, Berry G. Statistical methods in medical research. 2nd edn. Oxford: Blackwell, 1987;209.

15. Neave HR, Worthington PL. Distribution-free tests. London: Unwin Hyman Ltd, 1988;244-60.

16. Dobbs RJ, Lubel DD, Charlett A, et al. Hypothesis: age associated changes in gait represent, in part, a tendency towards parkinsonism. Age Ageing 1992;21:221-5.

17. Dobbs RJ, Charlett A, Bowes SG, et al. Is this walk normal? Age Ageing 1993;22:27-30.

18. O'Neill CJA, Weller C, Walker C, et al. Towards screening for a pre-clinical state in idiopathic parkinsonism: methodology. Eur J Clin Pharmacol 1989;36:140.

19. Bowes SG, Charlett A, Dobbs RJ, et al. Quantitative description of gait in idiopathic parkinsonism and of influencing factors. Scand J Rehabil Med 1992;24:(in press).

20. Visser H. Gait and balance in senile dementia of Alzheimer's type. Age Ageing 1983;12:296-301.

21. Sloman L, Berridge M, Homatides S, et al. Gait patterns of depressed patients and normal subjects. Am J Psychiatry 1982;139:94-7.

22. Calne DB, Langston JW. Aetiology of Parkinson's disease. Lancet 1983;ii:1457-9.

23. Barbeau A, Roy M, Paris S, et al. Ecogenetics of Parkinson's disease: 4-hydroxylation of debrisoquine. Lancet 1985;ii:1213-16.

24. Marttila RJ, Rinne E. Epidemiological approaches to the etiology of Parkinson's disease. Acta Neurol Scand 1989;80(suppl 126):13-18.

25. Todes CJ, Lees AJ. The pre-morbid personality of patients with Parkinson's disease. J. Neurol Neurosurg Psychiatry 1985;48:97-100.

26. American Academy of Orthopaedic Surgeons. Atlas of orthotics. Saint Louis: C.V. Mosby Co., 1975;260-1.

COPYRIGHT 1993 Oxford University Press
COPYRIGHT 2004 Gale Group