AI finds smoking spurs biological clock

Researchers at the University of Lethbridge used AI to analyze blood biochemistry in smokers—and found smoking makes people biologically older. Results of the study were published in Scientific Reports.

“By employing age-prediction models developed using supervised deep learning techniques, we found that smokers exhibited higher aging rates than non-smokers. In other words, we show that smoking makes people biologically older,” Olga Kovalchuk, MD, PhD, a professor in the Department of Biological Sciences at the University of Lethbridge in Canada, said in a prepared statement.

Kovalchuk and colleagues analyzed data from 149,000 anonymous individual blood biochemistry records linked to smoking status—49,000 individuals in the cohort were smokers. A machine learning algorithm looked for 66 different biomarkers in the blood that are associated with aging—including hemoglobin A1c (hbA1c), blood urea, fasting serum glucose and serum ferritin.

The machine learning algorithm, or “aging clock,” then guessed the cohort’s age based on the biomarker patterns in their blood. There was a linear association between blood markers and age, the researchers noted in their study. In fact, age prediction shows that the biological age of male smokers was 1.5 times older than their chronological age, while female smokers were nearly twice as old as their actual chronological age.

“Our study also demonstrated that young smokers (less than 40 years of age) have biological ages that are significantly higher than their chronological ages,” the researchers wrote in their study. “Surprisingly, this effect disappears in the oldest subjects.”

Deep learning-based hematological aging clocks can serve as “reasonably accurate predictors” of age for relatively healthy individuals and can also serve as accurate tools for evaluating the effect of lifestyle factors on biological aging. 

Additionally, they can act as distinguishers of patient smoking status—in other words, classifiers based on deep neural networks have the potential to support or even replace patient self-reporting and can “provide a better statistical assessment of the prevalence of tobacco smoking,” Kovalchuk and colleagues wrote in their study.

“The deep learning–based approach used in this study may be extended to analyze the combined effects of tobacco smoking and biochemically-defined diabetes mellitus and dyslipidemia as well as other potential morbidities,” the researchers wrote. “Similarly, DNNs could be used to predict health trajectories and outcomes or to evaluate the extent to which various other environmental exposures, dietary factors and genetic risks affect health and aging.”