Recently, increased rates of suicide in US counties at higher altitudes have been noted. Because of the documented association between cocaine use and suicide, we hypothesized that there would be a correlation between incidence of cocaine use and altitude of residence.
Cocaine use data were obtained from the Substate Substance Abuse Estimates from the 1999–2001 National Surveys on Drug Use and Health. Data related to the percentages of people 12 years or older who used cocaine in the past year. Average elevation for US counties was calculated using the Shuttle Radar Topography Mission elevation data set, and subject region elevation was calculated by averaging the weighted elevations of each region's relevant counties. The correlation between elevation of a substate region and incidence of cocaine use in that region was calculated using Pearson correlation coefficients.
A significant correlation exists between mean altitude of a substate region and incidence of cocaine use in that region (r = 0.34; P < 0.0001). Regression analysis controlling for age, sex, race, education level, income, unemployment, and population density was performed. Altitude remained a significant factor (P = 0.007), whereas male sex (P = 0.008) and possessing less than a college education (P < 0.0001) were also significant predictors of self-reported cocaine use in the past year. It is important to note that cocaine use was assessed in isolation of other drugs of abuse, an additional confounding variable.
This study demonstrates a significant correlation between altitude of substate region of residence and incidence of cocaine use. It is possible that stress response due to hypoxia is responsible; however, this requires further investigation. However, because other substance use was not assessed, specificity of this association is unknown. In addition, this correlation may help explain the increased rate of suicide in areas of higher elevation.
From The Brain Institute (KKF, DGK, PFR), University of Utah, Salt Lake City; Department of Radiology (NK), University of Ulsan College of Medicine, Ulsan, South Korea; Department of Psychiatry (DGK, PFR), University of Utah School of Medicine; and VISN-19 MIRECC (DGK, PFR), Salt Lake City Veterans Affairs Health Care System, Salt Lake City, UT.
Send correspondence and reprint requests to Kristen Fiedler, BS, The Brain Institute, University of Utah, 383 Colorow Dr, Room 347B, Salt Lake City, UT 84108. E-mail: email@example.com.
Supported by National Institutes of Health grants DA015116 and DA031247, VISN 19 MIRECC, and the Utah Science Technology and Research (USTAR) initiative.
The authors declare no conflicts of interest.
Received June 30, 2010
Accepted June 16, 2012