Abstract
Objective: To investigate predictors of survival in Parkinson disease (PD).
Methods: Vital status was determined in 800 subjects enrolled in a clinical trial of deprenyl (selegiline) and tocopherol 13 years earlier.
Results: Two hundred ninety-six deaths were recorded. There was no difference in the standardized mortality ratios across gender or age group. In univariate analyses, PD-specific variables associated with mortality were increased symmetry of parkinsonism, gait dysfunction as an initial symptom, severity of parkinsonism, and rate of worsening of parkinsonism prior to study enrollment. Cumulative exposure to deprenyl was not associated with mortality. In multivariable analysis, severity of parkinsonism and rate of worsening of parkinsonism remained associated with mortality. A poorer response to levodopa was associated with increased mortality independent of disease severity or dosage of levodopa. Results were unchanged when the analysis was restricted to 747 subjects maintaining a most likely diagnosis of PD throughout 6 years of active follow-up.
Conclusions: Parkinson disease did not affect survival differently across gender or age groups in this selected group of otherwise healthy clinical trial participants. Severity and rate of worsening of parkinsonism and response to levodopa are strongly related to survival.
Multiple studies have found that survival in Parkinson disease (PD) is reduced compared with the general population of the same age, including recent studies.1–3⇓⇓ The factors that determine survival in PD are not well established, however. Long median survival of patients with PD and the fact that factors unrelated to PD likely determine much of the variability in survival make this a difficult question to study. Many of the prospective studies that have been performed are limited by small study cohorts and relatively short duration of follow-up, resulting in small numbers of observed deaths, usually fewer than 100. Many studies are also limited by lack of follow-up information reaffirming the initial diagnosis of PD, resulting in uncertainty as to the frequency of inclusion of subjects with atypical parkinsonian syndromes. These atypical syndromes may have different determinants of survival than idiopathic PD.
We undertook a study to determine the factors that are associated with survival in PD in a large cohort of subjects with long-term follow-up. Clinical diagnostic information from the first several years of follow-up was used to allow more accurate diagnostic classification than in many previous studies.
Methods.
Subjects and data collection.
Clinical and demographic data were obtained from the Deprenyl and Tocopherol Antioxidative Therapy of Parkinson’s Disease (DATATOP) clinical trial database. The primary aim of the DATATOP trial was to investigate the impact of deprenyl (selegiline) and α-tocopherol on the progression of disability.4 Eight hundred subjects without severe postural instability (PD at Hoehn and Yahr stage5 1 or 2) within 5 years of symptom onset and not yet requiring symptomatic therapy were enrolled between September 1987 and November 1988. Subjects with important comorbid illness, cognitive impairment (Mini-Mental State Examination [MMSE] score of <23), or severe tremor (tremor score of ≥3 on the Unified Parkinson’s Disease Rating Scale6 [UPDRS]) were excluded. A comprehensive neurologic evaluation was performed at baseline as well as a general medical history including whether or not comorbid illness was present in each of 11 organ systems (yes or no by history only). Neither the specific comorbid illness nor its severity or significance was recorded. Data collected at follow-up visits included whether or not the subject had met the primary endpoint of sufficient disability to require treatment with levodopa and UPDRS mental, activities of daily living (ADL), and motor subscale scores. The mental and ADL subscales rate disabilities and impairments from the patient’s point of view. The motor subscale documents the severity of motor impairments rated by the clinician. The total score is the sum of scores on the three subscales. At each visit, the enrolling investigator was asked to state his/her level of confidence (0 to 100%) in a diagnosis of idiopathic PD and to rank up to three most likely diagnoses. We defined a subgroup of subjects maintaining a clinical course compatible with idiopathic PD as those for whom the investigator’s most likely diagnosis was PD with a level of confidence of at least 60% across all visits.
Potential prognostic factors.
A list of potential prognostic factors for survival was generated, including factors shown in previous studies to have predictive value. The prespecified factors were age, gender, pack-years of smoking, whether or not gait disorder was an initial symptom, whether or not tremor was an initial symptom, initial MMSE score, duration of symptoms prior to enrollment (according to the subject), baseline UPDRS scores (ADL and motor), prestudy rate of change of UPDRS scores, symmetry of parkinsonism, postural instability/gait disorder score, presence of an upgoing plantar response (present, absent, or equivocal), Hoehn and Yahr stage, cumulative exposure to deprenyl, and response to levodopa. Prestudy rate of change of UPDRS score was approximated by dividing the baseline total UPDRS score by the symptom duration recalled by the patient at enrollment. The symptom duration as assessed by the rater was also available and was found to be highly correlated with the duration estimated by the subjects (r = 0.92). A measure of symmetry of parkinsonism was calculated as follows: 
where right and left totals are the total of tremor, rigidity, and bradykinesia (manual dexterity and foot agility) items from the UPDRS for the right and left extremities. A postural instability/gait disorder score was calculated as the sum of the UPDRS ADL items “falling,” “freezing,” and “walking” and the UPDRS motor items “gait” and “postural instability.” Response to levodopa was defined as the percentage improvement in UPDRS motor scores with levodopa, calculated as the difference between the best score within 6 months after starting levodopa and the last pretreatment score multiplied by 100 and divided by the last pretreatment score.
Vital status ascertainment.
Vital status and dates of death were collected over the years 2001 to 2002. Data were obtained for the 694 US subjects using the Social Security Death Index (SSDI) and for the 106 Canadian subjects through contacting each enrolling investigator. When only the month and year of death were known, the date of death was recorded as the 15th day of the month. When only the year of death was known, the date of death was recorded as June 30 of that year. In 1997, contact with enrolling investigators and a finder’s service had been used to confirm vital status in 774 of the 800 subjects (26 individuals were lost to follow-up).7 For a death to be entered in the SSDI, the event must be reported to the Social Security Administration by families or funeral homes, and this is not always done. The updated search (2001 to 2002) was performed blind to the results of the 1997 ascertainment to allow the accuracy and sensitivity of the SSDI search to be assessed.
Statistical analysis.
Age-standardized mortality ratios (SMRs; observed/expected deaths) were calculated for men and women. SMRs were also calculated for subjects younger and older than the median age (62 years), standardized for 5-year age groups and gender. Mortality rates were standardized using gender-specific mortality tables published by the National Center for Health Statistics (USA) or Statistics Canada. Published death rates for 1995 (US) and 1997 (Canada) were used, as these represent the closest available years to the midpoint of the total follow-up period. CIs for the SMRs were calculated as described previously.8 The SMRs for younger vs older subjects and for men vs women were compared by testing the null hypothesis that their ratio is equal to 1 using an approximate χ2 statistic.9
The relationship between each of the variables of interest and time from study enrollment until death was investigated using Cox proportional hazards regression models. For subjects not identified as deceased, observation time was censored at the time of the SSDI search for US subjects and at the time of last clinic or telephone contact for Canadian subjects. Baseline variables that individually were significantly associated with time to death at the α = 0.2 level were entered into a multivariable model. Variables were then successively removed from the model if they were not significant at the α = 0.05 level and remained excluded if no evidence for interactions with other variables in the model was found as judged by changes in parameter estimates and significance levels of the remaining variables. A limited number of prespecified interactions (see appendix E-2 on the Neurology Web site at www.neurology.org) were also tested for significance in the resulting model. The proportional hazards assumption was tested by adding time-dependent covariates to the model and by estimating the hazard ratios (HRs) for each predictor variable over four postenrollment time intervals that contained approximately equal numbers of deaths. A test for homogeneity of these HRs was then carried out for each predictor variable. To assess model fit, predicted numbers of deaths within each quintile of risk defined by the model were calculated, and observed and predicted numbers of deaths were compared as described.10 The relationship between survival and prestudy rate of progression of parkinsonism was studied both in the whole cohort and in the subgroup of subjects with tremor as an initial symptom, because patients with other (more subtle or less well-recognized) presenting symptoms may underestimate the duration of their symptoms and therefore overestimate the rate of worsening of parkinsonism.
Response to levodopa and cumulative deprenyl exposure were not included in the model-building process as they are not baseline variables. These were investigated in separate multiple regression analyses. To estimate the effect of cumulative deprenyl exposure on survival, two time-dependent variables were used: cumulative deprenyl exposure and cumulative time on study. This was done to account for the fact that the time off-study, during which deprenyl exposure was unknown, varied from subject to subject.7 Deprenyl exposure in DATATOP occurred in three phases: First, subjects were randomized to receive either deprenyl or placebo. Second, after 14 ± 6 months of randomized assignment, all study subjects were offered open-label deprenyl. Third, approximately 4.5 years after the initial enrollment, 368 consenting subjects who had met the endpoint of sufficient disability to require levodopa were rerandomized independently of their initial blinded treatment assignment to receive either deprenyl or placebo.7 Deprenyl exposure was assigned according to intention to treat for the randomized phases and according to actual treatment recorded at each visit for the open-label phases. Compliance was judged to be very high (>97%) for the randomized phases of the study.11
Results.
At baseline, the mean age was 61 ± 10 years. Sixty-seven percent of the cohort were men. Median duration of PD at enrollment was 25 months (quartiles 15, 36), the mean total UPDRS score was 25 ± 12 points, and the mean motor UPDRS score was 17 ± 9 points. Other baseline characteristics of the cohort have been published in detail.4 The cohort was followed with in-person visits for a median of 6.5 years (quartiles 2.0, 7.1).
Vital status update.
Two hundred ninety-six subjects, or 37% of the cohort, were identified as deceased at the time of the 2001 to 2002 vital status update. The shortest time elapsed from enrollment to the vital status update was 13 years. One hundred fourteen of the 120 US subjects previously known to be deceased were identified as deceased by the new death index search (95% sensitivity). The remaining six subjects were classified as deceased for the purpose of this analysis. Of the 66 Canadian subjects not identified as deceased at the time of the present ascertainment, 50 had been contacted since January 1, 2000, and 35 since January 1, 2002.
SMRs.
A total of 9,453 person-years of observation were available for analysis. The distribution of observed and expected deaths and SMRs by age group and gender is shown in table 1. The death rate was similar to that expected for the general population (SMR = 1.04; 95% CI 0.93 to 1.17). When education-specific mortality rates were used (available only for the USA and only for individuals under age 64), the SMR in the under-64 US subgroup was 1.04 (95% CI 0.52 to 1.86) and the overall SMR rose slightly to 1.07 (95% CI 0.95 to 1.20). There were no differences between the SMRs of men and women or between older and younger subjects. This indicates that in this generally healthy cohort, PD did not have a differential effect on survival in men compared with women or in older compared with younger individuals.
Table 1 Standardized mortality ratios (SMRs) by age group and gender
Evolution of diagnosis and survival.
Of the 800 subjects, 747 retained a most likely diagnosis of PD with a degree of confidence throughout follow-up of >60% in the opinion of the enrolling investigator. Four other subjects retained a most likely diagnosis of PD across all visits but with a level of confidence of <60% at at least one visit. The most common diagnoses in the remaining 49 individuals were multiple-system atrophy (12), atypical parkinsonism of uncertain type (11), progressive supranuclear palsy (8), and diffuse Lewy body disease (4). Mortality was significantly greater in those subjects for whom the investigator did not indicate a most likely diagnosis of PD with a level of confidence of at least 60% throughout the period of follow-up (figure 1; p = 0.0004)
Figure 1. Kaplan-Meier curves of time to death stratified by retention of a high probability of a diagnosis of Parkinson disease (PD). Retained a >60% likelihood of PD: no (dashed line); yes (solid line). Log rank test, p = 0.0004.
All subsequent analyses were performed both in the whole cohort (n = 800) and in the subset of subjects who retained a >60% likelihood of a diagnosis of PD throughout follow-up (n = 747). The results did not differ significantly; therefore, only the results for the whole cohort are reported.
Characteristics associated with time to death.
Univariate analyses.
The HRs from the univariate analyses of characteristics at enrollment are shown in table 2. A number of baseline characteristics not specific to PD were associated with increased mortality, including older age, male gender, more pack-years of smoking, and lower MMSE scores. Among the comorbid illnesses, cardiac, pulmonary, and urologic comorbidities were associated with increased mortality. Being employed full-time was associated with better survival.
Table 2 Associations between baseline variables of interest and time to death: Univariate analyses
PD-specific variables associated with increased mortality were increased symmetry of parkinsonism, presence and severity of gait dysfunction, and several indexes of severity of parkinsonism (higher Hoehn and Yahr stage, increased UPDRS total, motor, and ADL scores). The rate of worsening of parkinsonism prior to the baseline evaluation was also positively associated with mortality, both in the whole cohort and in the subgroup of subjects with tremor as an initial symptom (HR = 1.3; p < 0.0001 for both analyses).
We did not find duration of disease or having tremor as an initial symptom to be significantly associated with mortality. Similarly, cumulative exposure to deprenyl did not have a significant association with mortality (HR = 1.0 for 1 year of additional exposure; 95% CI 0.9 to 1.1; p = 0.7). The presence of an extensor plantar response was not significantly associated with mortality, although the HR estimates for the presence of an equivocal or an extensor plantar response (relative to flexor responses on both sides) were both >1 (see table 2). In a post-hoc analysis, combining equivocal and extensor plantar responses yielded a significant association with greater mortality compared with flexor plantar responses on both sides (HR = 1.35; p = 0.03).
Multiple regression analysis.
The multiple regression model is shown in table 3. Comparing the observed with the model-based predicted numbers of deaths within each quintile of risk defined by the model did not reveal any strata for which the numbers differed, suggesting that the model fit the observed data well (data not shown).
Table 3 Multivariable extended Cox model for time to death
Figure 2. Hazard ratios associated with a 10-year increment in age at enrollment according to Hoehn and Yahr stage at enrollment and years of follow-up. (Diamonds) = stage 1; (squares) = stage 1.5; (triangles) = stage 2; (circles) = stage 2.5.
Figure 3. Hazard ratios associated with a one-level increase in Hoehn and Yahr stage by age at enrollment.
Several baseline characteristics not specific to PD remained associated with mortality (see table 3). The PD-specific characteristics that were independently associated with mortality in the multivariable analysis were Hoehn and Yahr stage, baseline total UPDRS score, and rate of worsening of parkinsonism prior to study enrollment. A time-dependent covariate was introduced, representing the interaction between time from enrollment and age on the risk of death, as there was evidence that the proportional hazards assumption was not met for the variable age (figure 2). An interaction was also found between Hoehn and Yahr stage and age. Older age was associated with increased mortality, and this effect was more marked at low Hoehn and Yahr stages than at higher Hoehn and Yahr stages (see figure 2). A higher Hoehn and Yahr stage was associated with significantly greater mortality for subjects who were younger at the time of enrollment, and this effect diminished with increasing age at enrollment (figure 3).
Response to levodopa as predictor of mortality.
Of the original 800 subjects, 498 had UPDRS scores available within the 6-month period after starting levodopa, allowing response to levodopa to be calculated as defined above. Many subjects did not have levodopa response data because initially no protocol was in place to follow subjects beyond the point of requiring levodopa. Of the 302 subjects without adequate UPDRS data, 110 subjects reached the endpoint of disability sufficient to require levodopa within 1 year, 49 terminated from the study within 1 year, 81 did not reach endpoint before 3 years of observation, and 62 reached the endpoint or terminated between 1 and 3 years after enrollment. Thus, they displayed a wide range of outcome with respect to progression of motor disability. The 498 subjects included in the analysis did not differ from the 302 excluded subjects with respect to age, gender, UPDRS score, or Hoehn and Yahr stage at baseline; however, they had a faster rate of change of UPDRS scores prior to requiring levodopa (9 points/year [quartiles 5, 15] vs 6 points/year [quartiles 2, 17]; p = 0.004). A higher proportion maintained at least a 60% likelihood of a diagnosis of PD throughout their period of observation (95 vs 91%; p = 0.04), and they also had slightly better survival.
UPDRS scores were available a median of 62 days (quartiles 39, 101) prior to the first visit on levodopa. The minimum motor UPDRS score occurred a median of 28 days (quartiles 0, 98) after the first visit on levodopa. The average response to levodopa was an improvement of 29 ± 30% on the UPDRS motor score relative to the pretreatment score.
In a model including the UPDRS motor score at the last visit prior to starting levodopa, disease duration, and all baseline factors found to be associated with survival in the multivariable analysis, a 20% greater response to levodopa was associated with better survival (HR = 0.74; 95% CI 0.70 to 0.80; p < 0.0001). This result suggests that independent of disease duration, severity, age, gender, and other relevant factors, a person responding well to levodopa will have a longer survival than an individual responding poorly. When the analysis was also adjusted for the dosage of levodopa at the time of the best recorded motor UPDRS score within 6 months of beginning levodopa, a 20% better response to levodopa was still associated with better survival (HR = 0.72; 95% CI 0.67 to 0.77; p < 0.0001). This suggests that the association between survival and a greater improvement in UPDRS scores with levodopa does not simply reflect the intensity of treatment with levodopa. Within the subgroup of individuals for whom the investigator maintained a level of confidence of at least 60% for a diagnosis of PD across all visits, the HR associated with a 20% better response to levodopa was 0.78 (95% CI 0.72 to 0.84; p = 0.001).
Discussion.
Our analysis has estimated mortality rates and identified associations with mortality in a clinical trial population. Mortality rates in this cohort will be affected by numerous selection factors, both known and unknown, which operate to make clinical trial cohorts different from the general population. There is evidence that participants in research are different from eligible nonparticipants in ways that favorably influence their survival.12,13⇓ As well, the health of clinical trial participants may benefit directly from being part of the trial. This distinction between the DATATOP subjects and the general population must be kept in mind when interpreting the results.
The mortality rate in this cohort was not different from that of the general population, which contrasts with the findings of many other investigators.1–3⇓⇓ When investigators have found mortality rates no different from those of the general population,14–16⇓⇓ proposed explanations have included delays in deaths due to the introduction of levodopa17 and cohorts of patients with milder, earlier disease.18 The lower SMR in the DATATOP cohort compared with most other reports of mortality in PD likely reflects, at least in part, the exclusion of subjects with unstable medical illness from the cohort at enrollment. Seven percent of the 571 potential subjects screened but determined to be ineligible for the study were excluded on the basis of unstable medical illness.4 A previous analysis of the DATATOP cohort found a mortality rate less than that expected in the general population.7 At the time of the previous investigation, less time had elapsed since enrollment so the selection effects of the enrollment criteria would have been even more marked. Subsequently, as this selection effect lessened, an excess of deaths relative to the general population must have occurred to bring the overall mortality rate up to that of the general population. Repeated analysis of mortality rates in other cohorts has also shown increasing SMRs over time, suggesting that advanced PD has a more negative effect on survival than early PD relative to the general population.2,3,19,20⇓⇓⇓ The DATATOP cohort was also more highly educated than the general population (689/800 subjects had completed ≥12 years of formal education). The socioeconomic status of the cohort is likely also above average. Age- and education-standardized SMRs are available only for the US population and only under the age of 64; therefore, our education standardization is incomplete, and we were not able to account for economic status. These factors will bias the result toward a lower overall SMR, as educational level and socioeconomic status are inversely associated with mortality.
The age- and sex-standardized SMRs for men and women were not found to be different, suggesting that PD has a similar effect on the survival of men and women. Several previous studies have not found differences in survival between men and women with PD relative to control subjects or to the general population.2,14,21,22⇓⇓⇓ Studies that have found differences have found women to have 1.5 to 4 times higher ratios of observed to expected deaths than men, when gender-specific death rates are used for standardization.5,23,24⇓⇓ We observed a slight trend toward a higher SMR in women. Because rates of serious medical illness are higher in older men than in older women, excluding such patients would result in a greater underestimation of the true mortality rate in men than in women. This reasoning suggests that in a less selected cohort, a significantly higher SMR in women than in men would not be found. In three studies of unselected community-based cohorts comparing mortality rates in men vs women while adjusting for gender-specific mortality rates in a control population without PD, no significant differences were found.22,25,26⇓⇓ Our results are consistent with these findings.
We found that survival compared with the general population was not different between older and younger individuals. This result is consistent with a study of mortality rates in young-onset PD (21 to 40 years), which found an SMR of 2.1, similar to SMRs found in PD cohorts not selected for age at onset.27 However, older individuals selected for good general medical health would likely be more different from the general population of their age than younger individuals. This differential selection could mask a greater negative effect of PD on survival in the older subjects in the cohort, resulting in their SMR appearing to be no different from that in the younger subjects. Three previous studies have found higher than expected mortality in older individuals with PD compared with younger individuals.19,21,22⇓⇓ None of these studies selected participants for good general medical health, nor was there adjustment for comorbid illnesses. A possible explanation for the contrast between our findings and these three studies is that PD interacts with other medical illnesses to reduce survival.
We found that a number of clinical and demographic factors not specific to PD were strongly associated with mortality. This is in keeping with the observation that patients often die of causes unrelated to PD. Of note is our finding of an association between greater pack-years of smoking and increased mortality. A recent case-control study found that a past history of smoking was not associated with increased mortality in the PD patients, whereas it was a significant marker of increased mortality in control subjects.25 Their finding suggests a possible protective effect of smoking in PD patients. A limitation of our study is the lack of an age-matched control group to provide an estimate of the expected magnitude of the effect of pack-years of smoking on time to death.
The association between equivocal or extensor plantar responses and increased mortality could reflect either individuals with atypical parkinsonism such as multiple-system atrophy in the cohort or neurologic comorbidity such as cerebrovascular disease reducing survival. The effect of abnormal plantar responses on survival was similar in the group for which the investigator was relatively confident of a diagnosis of PD throughout follow-up. In the 49 individuals with alternative clinical diagnoses, abnormal plantar responses were not more common than in the rest of the cohort (19% for both subgroups). Both of these observations suggest that neurologic comorbidity unrelated to parkinsonism may be responsible for reduced survival in individuals with abnormal plantar responses.
We also identified associations with mortality related to the clinical features of PD, most importantly baseline severity of parkinsonism and rate of worsening of parkinsonism prior to the start of observation. Previously published studies have shown similar findings.2,21,28,29⇓⇓⇓ Rate of worsening of parkinsonism prior to enrollment was based on subject recall, and this could be a source of bias. For example, subjects with more subtle initial symptoms (such as loss of facial expression or reduced arm swing) may underestimate the duration of their symptoms relative to those beginning with tremor, resulting in an overestimate of the rate of worsening of parkinsonism prior to enrollment. However, an analysis confined to those subjects with tremor as an initial symptom was consistent with the findings in the whole cohort.
Hoehn and Yahr stage was associated with a more negative effect on survival at younger ages, suggesting that severity of parkinsonism is a more important prognostic factor in younger patients. In older individuals, comorbid illnesses that influence survival may dilute the importance of the severity of parkinsonism. Age appears to be a more important prognostic factor for survival in those with mild parkinsonism (low Hoehn and Yahr stage) at baseline. In those with more severe PD, the effect of parkinsonism on survival may dominate over age and concomitant medical illnesses.
Gait dysfunction and increased symmetry of parkinsonism were significantly associated with increased mortality in univariate analyses but not in the multivariable analysis. These associations were maintained in the subgroup maintaining a clinical course compatible with PD over the period of observation, suggesting that they are not due to including subjects with reduced survival due to atypical parkinsonism. We cannot rule out, however, that subjects with atypical parkinsonism are not included in the subgroup of individuals maintaining a clinical course compatible with PD.
Several previous studies have noted rest tremor at onset to be associated with slower progression of parkinsonism.30 We did not find an association between mild tremor at onset and survival; however, subjects with severe tremor at baseline were excluded from the cohort. Investigating the prognostic importance of disease duration and cognitive impairment was similarly compromised because the DATATOP cohort was restricted to subjects with <5 years of symptom duration at study entry and with MMSE scores above 22.
We did not find any association between cumulative deprenyl use and mortality. Deprenyl’s effect on mortality was of interest because of the proposed neuroprotective action of this compound and because of the observation that mortality was increased in participants of a clinical trial of PD who had been randomized to receive deprenyl and levodopa compared with levodopa alone.31 This was followed by a population-based study in the United Kingdom finding a slight but nonsignificant excess mortality among deprenyl users.32 Our findings are consistent with a subsequent meta-analysis of five randomized trials, not including DATATOP, of deprenyl vs placebo in PD that did not detect any significant increase in mortality for those patients receiving deprenyl.33 Although our results are further suggestive that the results of the Parkinson Disease Research Group of the United Kingdom were the result of a type I error, we cannot provide evidence against an alternative hypothesis raised by the authors that a subgroup (e.g., those with particular medical comorbidities) are particularly susceptible to adverse effects of deprenyl alone or in combination with levodopa.34
We found that a poorer response to levodopa was a significant predictor of increased mortality independent of disease severity at the time of initiating treatment and independent of levodopa dosage. This association was as strong when the analysis was restricted to those maintaining a clinical course highly compatible with PD, indicating that it is unlikely that the association is due to including individuals with atypical parkinsonism characterized by poorer survival35–37⇓⇓ and poor response to levodopa.36,38⇓ Whether a poor response to levodopa marks a biologically distinct subgroup of PD with a poorer prognosis or retaining a worse parkinsonian state reduces survival remains to be clarified.
Acknowledgments
Supported in part by research fellowships from the Canadian Institutes of Health Research and the Parkinson Society Canada (C.M.), a Center of Excellence Grant from the National Parkinson Foundation (A.E.L.), and the Mary Trimmer Chair in Geriatric Medicine Research, University of Toronto (G.N). The DATATOP trial was supported by US Public Health Service grant NS 24778.
The authors thank the Canadian site investigators and coordinators for their assistance in updating vital status information on their patients.
Footnotes
-
Additional material related to this article can be found on the Neurology Web site. Go to www.neurology.org and scroll down the Table of Contents for the January 11 issue to find the title link for this article.
-
↵*See appendix E-1 on the Neurology Web site for a list of Group members.
- Received March 16, 2004.
- Accepted September 15, 2004.
References
Letters: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
You May Also be Interested in
Hemiplegic Migraine Associated With PRRT2 Variations A Clinical and Genetic Study
Dr. Robert Shapiro and Dr. Amynah Pradhan