Dopamine agonist–induced impulse control disorders may be more likely in Parkinson’s disease patients who smoke, drink, or consume caffeine, according to the first prospective cohort study of the problem.

“We found a number of interesting things” when the 18 patients (39%) in the study who developed an impulse control disorder (ICD) were compared with the 28 who did not, said lead author Jesse Bastiaens, a medical student at Cornell University, New York.

The 18 had a higher baseline prevalence of caffeine use (100% vs. 67%; P = .007), lifetime prevalence of cigarette smoking (44% vs. 14%; P = .04), and cumulative exposure to both caffeine (72 vs. 38 cup-years; P = .04) and cigarettes (median 2 vs. 0 pack-years; P = .07). There was also a nonsignificant trend towards a higher baseline prevalence of alcohol use.

At baseline, ICD patients also had a greater prevalence of motor complications (61% vs. 25%; P = .01) and lower baseline modified Hoehn and Yahr Scale scores (mean 1.6 vs. 1.9; P = .05), despite comparable total Parkinson’s disease (PD) drug use in both groups (median, 150 levodopa equivalents in both groups).

ICD patients had higher peak-dopamine agonist (DA) doses (median, 300 vs. 165 levodopa equivalents; P = .03) but no significant differences in DA treatment duration or cumulative exposure.

“The risk of ICDs in PD is related to both patient-specific factors and peak DA dosage. Attention to these susceptibility factors may help to reduce the incidence of ICDs in PD,” Mr. Bastiaens said at the annual meeting of the American Academy of Neurology.

Besides counseling patients about the risk, “you might be less eager to use a dopamine agonists or use lower doses in people with prior histories of smoking or heavy drinking or other addictive behaviors,” said Dr. Ronald Pfeiffer, vice chair of the department of neurology at the University of Tennessee, Memphis, who moderated Mr. Bastiaens’ presentation.

Patients who developed an impulse disorder did so after a mean treatment duration of 23 months (range, 3-114 months). Age, age of PD onset, sex, depression, anxiety, and other factors were not predictive.

None of the 46 patients (all outpatients) had previous histories of ICDs, and none were demented. Their mean age was 62 years, and about half were women. ICD was diagnosed by semistructured interviews and clinical data, with criteria that were perhaps a bit less strict than were those in the DSM-IV, according to Mr. Bastiaens, who said that the researchers thought that casting a slightly wider net might be useful.

There was 1 case per 100 person-months of DA exposure. “We found that impulse control disorders can really occur at anytime during dopamine agonist therapy – as early as 3 months or 9 or more years,” said Mr. Bastiaens, who noted that a questionnaire to assess for ICDs in Parkinson’s patients has been recently validated (Mov. Disord. 2012;27:242-7).

A 2010 cross-sectional study in 3,090 Parkinson’s patients found an association between ICDs and current cigarette smoking and family histories of gambling problems, among other factors. It did not find an association with DA dosage. Just over 17% of patients on the drugs developed ICDs (Arch. Neurol. 2010;67:589-95).

The study was supported by the Parkinson’s Disease Foundation. Mr. Bastiaens said that he had nothing to disclose. Dr. Pfeiffer reported personal compensation from several pharmaceutical companies, none of which were involved in the study.