Intended for healthcare professionals


Occurrence of different cancers in patients with Parkinson's disease

BMJ 1995; 310 doi: (Published 10 June 1995) Cite this as: BMJ 1995;310:1500
  1. Henrik Moller, epidemiologista,
  2. Lene Mellemkjaer, epidemiologistb,
  3. Joseph K McLaughlin, epidemiologistc,
  4. Jorgen H Olsen, epidemiologistb
  1. a International Agency for Research on Cancer, F-69372 Lyons, France
  2. b Danish Cancer Society, Division for Cancer Epidemiology, Copenhagen, Denmark
  3. c National Cancer Institute, Epidemiology and Biostatistics Program, Bethesda, Maryland, USA
  1. Correspondence to: Dr Moller.
  • Accepted 21 March 1995

Parkinson's disease is a chronic, progressive neurological disorder with well defined clinical and pathological features. Smokers are less likely to develop Parkinson's disease.1 The disease is associated with a reduced risk of cancer overall,2 but the risk of malignant melanoma may be increased.3 The objective of this study was to assess the occurrence of cancer in a large cohort of patients with Parkinson's disease.

Patients, methods, and results

The study was based on three computerised registers in Denmark. From the Danish hospital discharge register we extracted all records of admissions of patients with a primary diagnosis of Parkinson's disease during 1977-89 and identified a cohort of 7046 people with the disease (3470 men and 3576 women). Information on cancer incidence and death among cohort members from their first recorded admission for Parkinson's disease till the end of 1990 was obtained from the Danish cancer registry and from the Danish register of deaths. The expected numbers of cases of cancer were calculated from the person years at risk among cohort members and the incidences of cancer in the Danish population, with due account being taken of age and calendar period. Relative risks were calculated as the ratios of the observed to the expected numbers of cases of cancer and are presented with their 95% confidence intervals. The average duration of follow up was 4.6 years.

The overall incidence of cancer was lower than expected (table; relative risk 0.88). Relative risks were significantly reduced for lung cancer (0.29) and bladder cancer (0.42). Significantly increased relative risks were seen for skin melanoma (1.96) and for malignancies of unknown origin (1.76). Brain tumours occurred more frequently than expected (relative risk 1.61). The excess was, however, confined to the first three years after Parkinson's disease had been diagnosed (11 cases; relative risk 2.42 (95% confidence interval 1.2 to 4.3)). The combined relative risk of cancers that are known to be associated with tobacco smoking was reduced (0.49), but the combined relative risk of cancers of specified sites other than those associated with smoking was close to that expected (1.01).

Relative risks of cancer in 7046 people with a primary diagnosis of Parkinson's disease in Denmark, 1977-89

View this table:


We found a low incidence of cancers associated with tobacco smoking, mainly lung and bladder cancer, among patients with Parkinson's disease. In addition, smokers are less likely to develop Parkinson's disease.1 The most simple explanation of these findings is that tobacco smoking in some way exerts a protective effect against the development of Parkinson's disease, possibly through the effect of nicotine on dopamine concentrations in the brain,5 and that patients with Parkinson's disease therefore tend to have smoked less during their lives than the population at large.

The close temporal association between diagnosis of Parkinson's disease and incidence of brain tumours suggests that brain tumours may on rare occasions be erroneously diagnosed as Parkinson's disease. We found an excess of six cases of brain tumours among 7046 patients with Parkinson's disease. This implies that no more than one in 1000 diagnoses of Parkinson's disease is an unrecognised brain tumour.

The increased incidence of malignant melanoma lends support to the idea that treatment with levodopa may increase the incidence of malignant melanoma.3 Malignant melanoma cells possess a unique biochemical pathway for the conversion of levodopa to melamine, and they selectively incorporate levodopa.4 Levodopa treatment may therefore accelerate the growth of a preclinical melanoma. We emphasise, however, that the observed effect is rather weak: around one excess case per 3500 patients with Parkinson's disease per year.

We thank Dr Annie Sasco for helpful comments on a draft of the article.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract