ARTICLE IN BRIEF
Researchers presented new findings suggesting that nearly half of patients who receive an early diagnosis of Parkinson's disease (within 5 years of symptom onset) may not actually have the condition at autopsy.
How accurate is a clinical diagnosis of early Parkinson's disease (PD)? Not very, according to new research presented at the 2014 AAN Annual Meeting in Philadelphia.
Charles Adler, MD, PhD, FAAN, and colleagues reported that many of the patients who were newly diagnosed (less than five years) with Parkinson's based on clinical observations did not have confirmed PD when looked at during autopsy. The researchers based their findings on data from the Arizona Study of Aging and Neurodegenerative Disorders (AZSAND), an ongoing longitudinal clinical-neuropathological study, in order to determine the predictive value of an antemortem PD diagnosis,
“There are little data on this type of diagnostic accuracy for subjects with PD for less than five years, and that was the main point of our paper. The accuracy of diagnosis during life is tremendously critical to whether studies of living patients with PD are truly accurate,” Dr. Adler, professor of neurology at the Mayo Clinic College of Medicine in Arizona, told Neurology Today.
“While it is assumed that the longer one follows a PD patient, the more likely it is they have PD, this data is sobering in the sense that, even when medication responsive, cases that were seen with a PD duration of less than five years only had PD 53 percent of the time at autopsy,” he said. “Therefore, caution is needed when interpreting clinical studies of PD that do not have autopsy confirmation and reinforce the need for a tissue or other diagnostic biomarker,” the authors wrote.
Dr. Adler and colleagues in Arizona have enrolled participants from AZSAND with or without PD (controls) into the Brain and Body Donation Program (BBDP; brainandbodydonationprogram.org) since 1997. “We examine them annually and at each visit determine whether they likely have PD, including first visit. At the time of death we then review all our records as well as any outside medical records to determine whether we think the subject had PD or a different disorder. The only way to make a final diagnosis of PD is by doing and autopsy and the brain needs to have Lewy bodies and neuronal loss in the substantia nigra,” Dr. Adler told Neurology Today.
The two clinical diagnostic confidence levels used for the study presented at the Annual Meeting were probable PD (ProbPD – responsive to dopaminergic medications) and possible PD (PossPD – never treated or not clearly responsive), measured at two diagnostic time points, first visit and final (postmortem) conference. The neuropathological diagnosis was used as the gold standard for all calculations.
Among individuals diagnosed with ProbPD at first visit with disease duration of less than five years, the investigators reported that the diagnosis was confirmed neuropathologically in only about half of the cases (8/15). For individuals diagnosed with ProbPD with greater than five-year disease duration upon first visit (often presenting with motor fluctuations and/or dyskinesias), the diagnosis was confirmed neuropathologically in 88 percent (72/82).
Overall, the diagnosis of ProbPD at first visit in the 97 subjects was accurate 82 percent of the time (80/97). In those patients with PossPD at first visit, only 26 percent (9/34) had neuropathologically confirmed PD. For the final visit time point, 91 out of 107 patients with ProbPD had neuropathologically confirmed PD. Diagnostic accuracy improved with clinical variables such as response to levodopa, motor fluctuations, dyskinesias, and hyposmia, the authors wrote.
Importantly, “all subjects were seen by three movement disorder physicians, the number of probable PD cases was large, all were followed to autopsy with neuropathological confirmation,” said Dr. Adler. However, a limitation was that “subjects with possible PD were older (with a mean age of onset 80.6) than the age of our probable PD cases with mean age of onset 67,” and that the clinical exam findings that occur with aging may have complicated the diagnostic accuracy, he added.
Should these findings alter prescribing practices for patients with suspected PD? Not yet, said Dr. Adler, since most of the patients without confirmed PD upon autopsy “did have a different neurodegenerative disorder and since there are no different treatments available for these, it is appropriate that physicians continue to use the PD meds to try to improve symptoms and quality of life.”
These data should now be used to guide research in early PD, he said. “Studies on genetics, epidemiology, biomarkers, and treatments in early PD all need to be aware of the limitation that many of the subjects being entered may not actually have PD. It is critical that studies be powered correctly. It is also critical that these types of studies attempt to follow patients through to autopsy to obtain neuropathologic confirmation of the diagnosis until we find a biomarker that clearly has high sensitivity and specificity for PD. Our work, and that of other groups, investigating tissue biopsy for a diagnosis of PD during life will be essential to moving the field forward.”
This study was funded by National Institute of Neurological Disorders and Stroke, National Institute on Aging, Arizona Biomedical Research Commission, Michael J. Fox Foundation for Parkinson's Research, and Mayo Clinic Foundation.
Joseph Jankovic, MD, FAAN, professor of neurology, distinguished chair in movement disorders, and director of the Parkinson's Disease Center and Movement Disorders Clinic in the neurology department at the Baylor College of Medicine in Houston, TX, said that these data are important because they come from a prospective longitudinal study, unlike prior clinical-pathological studies. “Furthermore, the clinical assessments and diagnoses were made by seasoned Parkinson's disease specialists and the autopsies were performed by experienced neuropathologists.” And despite the seemingly “ideal conditions,” he said, only 53 percent of patients with less than five years disease duration had PD confirmed on autopsy.
The meeting abstract, he added, provides only limited information. “For example, it does not indicate what clinical and pathological diagnostic criteria were used. Nevertheless, the reported diagnostic accuracy is substantially lower than the 98.6 percent (72 of 73) positive predictive value of the clinical diagnosis of PD made by movement disorder neurologists in patients whose brains were collected at the United Kingdom Parkinson's Disease Society Brain Research Centre.” He noted, however, that the individuals in the UK study had much longer symptom duration before death than those in the AZSAND study. Another study of patients with early PD with very different findings, led by Dr. Jankovic, was the DATATOP study, which followed 800 individuals with clinically diagnosed PD for a mean of 7.6 years through an analysis of autopsy data, imaging studies, response to levodopa, and atypical clinical features, indicating a 91.9 percent accuracy of initial diagnosis of PD, but the final diagnosis was not based on pathologic confirmation in all cases.
In many instances, he said, “the disease does not fully declare itself clinically until patients are followed for more than several years” so the first time point in this study may have been too early to observe the clinical changes of PD.
Interestingly, Dr. Jankovic said, hyposmia was listed as a clinical variable that improved diagnostic accuracy along with response to levodopa and occurrence of levodopa-related motor complications, since several studies have now shown “that loss of smell can precede PD motor symptoms by several years or even decades.”
It was not clear from the abstract “whether the olfactory bulb was examined for evidence of synuclein pathology and whether those individuals who reported loss of sense of smell had more severe olfactory bulb neurodegeneration than those with normal olfaction,” said Dr. Jankovic. Also, he wondered what the final neuropathological diagnosis was in those individuals who did not meet the criteria for the clinical or pathological diagnosis of PD upon autopsy, and whether any other tools, such as DaTscan, were used to assist with the diagnosis.
Notably, he said, this study “draws attention to the importance of diagnostic vigilance when attempting to diagnose individuals with early parkinsonian features. The clinician must be willing to change the diagnosis, particularly with the evolution of atypical features, during subsequent follow-up.” Dr. Jankovic said he looks forward to the final publication of this work “and for additional information that may explanation why this study yielded such disappointing diagnostic accuracy.”
This study was selected by members of the Neurology Today editorial advisory board as one of the noteworthy papers from the AAN Annual Meeting.