Research Review Smoking and Parkinson’s Disease: Systematic Review of Prospective Studies Mohamed Farouk Allam, MPH, PhD, 1 * Michael J. Campbell, PhD, 2 Albert Hofman, PhD, 3 Amparo Serrano Del Castillo, MD, PhD, 1 and Rafael Ferna ´ndez-Crehuet Navajas, MD, PhD 1 1 Department of Preventive Medicine and Public Health, Faculty of Medicine, University of Cordoba, Cordoba, Spain 2 Institute of General Practice and Primary Care, University of Sheffield, Sheffield, United Kingdom 3 Department of Epidemiology and Biostatistics, Erasmus Medical Center Rotterdam, Rotterdam, The Netherlands Abstract: We estimated the pooled risk of tobacco smoking for Parkinson’s disease (PD). Inclusion criteria included systematic searches of MedLine, PsycLIT, Embase, Current Contents, previously published reviews, examination of cited reference sources, and personal contact and discussion with several in- vestigators expert in the field. Published prospective studies on PD and cigarette smoking. When two or more studies were based on an identical study, the study that principally investi- gated the relationship or the study that was published last was used. Seven prospective studies were carried out between 1959 and 1997, of which six reported risk estimates. Four cohorts were based on standardised mortality rates, which were exclu- sively of male. Only one study included risk estimates for both males and females separately. The risk of ever smoker was 0.51 (95% confidence interval, 0.43 to 0.61). There was an obvious protective effect of current smoking in the pooled estimate (relative risk, 0.35; 95% CI, 0.26 – 0.47). Former smokers had lower risk compared with never smokers (relative risk, 0.66; 95% CI, 0.49 – 0.88). Although our pooled estimates show that smoking is inversely associated with the risk of PD, the four prospective studies that were based on follow-up of mortality of smokers had many limitations. Further studies evaluating the association between smoking and PD in women are strongly needed. © 2004 Movement Disorder Society Key words: Parkinson’s disease; prospective study; smok- ing; systematic review It is now more than 40 years since Dorn (1959) published the first study suggesting an inverse association between cigarette smoking and Parkinson’s disease (PD) risk. 1 Ever since, this topic has been a controversial issue with the findings from several subsequent epidemiological studies. Many case– control studies have confirmed that, during the premorbid period, exposure to nicotine is lower in patients with PD than control subjects. 2–7 However, several other studies did not confirm the association. 8 –14 Tzourio and colleagues (1997), after pooling the results of five European population-based case– control studies, found no overall association be- tween cigarette smoking and PD. The authors reanal- ysed their results according to age of patients, and ever smoking was protective only in the lowest quartile age groups, which could be a spurious subgroup finding. 15 These inconsistent findings may be due to an etiolog- ical heterogeneity. However, methodological differ- ences, sample sizes, and regional differences in smok- ing habit may be the explanation. Morens and coworkers (1995) in their narrative review on the relationship between smoking and PD noted sys- tematic biases in many articles. Reported systematic biases were selective mortality, cause-and-effect bias, and smoking-associated symptom/sign suppression. 16 Systematic biases usually affect case– control rather than *Correspondence to: Dr. Mohamed Farouk Allam, Department of Preventive Medicine and Public Health, Faculty of Medicine, Univer- sity of Cordoba, Avda. Mene ´ndez Pidal, s/n Cordoba 14004, Spain. E-mail: fm2faahm@uco.es Received 28 May 2002; Revised 27 September 2003; Accepted 10 November 2003 Published online 12 February 2004 in Wiley InterScience (www. interscience.wiley.com). DOI: 10.1002/mds.20029 Movement Disorders Vol. 19, No. 6, 2004, pp. 614 – 621 © 2004 Movement Disorder Society 614