Does Parkinson’s disease occur more often in the relatives of patients with Parkinson’s disease, and if so, does the information provided by family history studies offer a reliable and accurate means of investigating the role of heredity in Parkinson’s disease? Two studies and an editorial published in the July 8 Neurology addressed these questions and asserted both the utility and the dangers of such studies.

In the first study, led by Walter A. Rocca, MD, MPH, Professor of Epidemiology and Neurology at the Mayo Clinic, Rochester, Minnesota, the family history method was compared with the family study method for the validity of information provided. Case subjects with Parkinson’s disease were found to be more aware of Parkinson’s disease among their first-degree relatives than control subjects. However, they overreported Parkinson’s disease in relatives who were not affected, leading to what the investigators termed “a substantial family information bias.”

In the second study, conducted by Karen Marder, MD, MPH, Professor of Neurology in the G. H. Sergievsky Center and Taub Institute for Alzheimer’s Disease and the Aging Brain, and colleagues at Columbia University, New York City, a family history interview (FHI) was created to ascertain the accuracy of data obtained in family history studies. An algorithm based on responses to the FHI was divided into four diagnostic categories and tested for sensitivity and specificity compared to the “gold standard” diagnosis of personal or medical record examination. “The FHI can be used to obtain reliable and valid family history information of Parkinson’s disease in first-degree relatives when a conservative diagnostic algorithm is applied,” the researchers reported.

FAMILY HISTORY OR FAMILY STUDY

Alexis Elbaz, MD, PhD, first author of the Mayo Clinic study, explained that the family history method involved collecting information on the occurrence of Parkinson’s disease among relatives by interviewing the case subjects and control subjects, or one proxy for each family. By contrast, the family study method required the study of each relative of the case and control subjects through a combination of direct and proxy interview, examination, or review of medical records. In the Mayo Clinic study, the family study method was considered the standard of comparison for the investigation of the validity of the family history method.

Of 202 case subjects and 202 age- and sex-matched control subjects, 133 population-based case subjects and their 655 relatives, and 119 control subjects and their 511 relatives were included in the study’s analysis. Interviews and screening for Parkinson’s disease were conducted by phone.

The family history method resulted in a 68% sensitivity for case subjects and 45% sensitivity for controls, reported Dr. Elbaz, a visiting scientist in the Department of Health Science Research, Mayo Clinic, while on leave from the INSERM Unit 360, Hôpital de la Salpêtrière in Paris. However, case subjects also had a greater overreporting of Parkinson’s disease in unaffected relatives: the specificity of the family history method was 99% for cases and 100% for controls, Dr. Elbaz noted. “A 1% difference in specificity caused a substantial difference in absolute numbers (one versus eight probands misclassified as positive family history) because the numbers of unaffected relatives were large,” he said.

The odds ratio for family history of Parkinson’s disease was 4.34 using the family history method and 1.86 using the family study method, indicating that there was significant bias in the family history method, resulting from the lower sensitivity of control subjects and the lower specificity of case subjects, according to the researchers. The authors concluded that “The risk of misclassifying relatives affected by another type of parkinsonism or by essential tremor as [having] Parkinson’s disease is substantial and should be considered when interpreting findings from the family history method.”

ACHIEVING ACCURACY?

The study by Dr. Marder and colleagues, which focused on the accuracy of their FHI, resulted in a more conservatively positive interpretation of the utility of the family history method.

The FHI was assessed on four algorithm-derived diagnostic categories using different levels of diagnostic certainty of Parkinson’s disease: “definite,” “definite or probable,” “conservative” (definite, probable, or possible), and “liberal” (definite, probable, possible, or uncertain), Dr. Marder explained. The sensitivity and specificity of each diagnostic category was analyzed against the gold standard of medical examination or medical record review.

In the validity analysis, “the conservative diagnosis provided the best combination of sensitivity (95.5%) and specificity (96.2%) for the proband’s family history report,” Dr. Marder said. This reliability study included 2,225 first-degree relatives (636 parents, 809 siblings, 780 offspring) of 304 probands with Parkinson’s disease and 232 controls. Histories were provided by both the proband and another informant. Results were compared between proband versus self-report, proband versus sibling of subject, and proband versus child of subject. “Agreement between informants was excellent for definite or probable Parkinson’s disease for all three pair-wise comparisons,” she added.

The researchers also noted that, because the information from cases and their relatives was similar, one family history informant seemed adequate. With the application of “specific criteria to derive categories of diagnostic certainty,” the FHI “can be used to collect reliable and valid information on the diagnosis of Parkinson’s disease in families,” they concluded.

CAVEAT HISTORIAN

In her accompanying editorial review of both the Mayo Clinic and the Columbia University studies, Caroline M. Tanner, MD, PhD, Director of Clinical Research at the Parkinson’s Institute in Sunnyvale, California, praised the work of the researchers but remarked that “caution is advised” in employing the family history method alone to investigate Parkinson’s disease. As the two studies’ results suggest, “the relative’s report alone is not always accurate. Whenever possible, diagnosis should be verified by examination, and, when not possible, by medical record review,” she emphasized, particularly when considering the question of genetic screening for patients with Parkinson’s disease.

NR

—C. Justin Romano

Suggested Reading
Elbaz A, McDonnell SK, Maraganore DM, et al. Validity of family history data on PD: evidence for a family information bias. Neurology. 2003;61:11-17.
Marder K, Levy G, Louis ED, et al. Accuracy of family history data on Parkinson’s disease. Neurology. 2003;61: 18-23.
Tanner CM. PD or not PD? That is the question. Neurology. 2003;61:5-6.

Return to table of contents