Obesity-cancer link gaining plausibility
Obesity during adolescence is associated with increased midlife cancer risk, according to findings from a large population-based cohort of Israeli teens examined between 1967 and 2010.
The association, which was stronger in individuals in the later period of the cohort than in those in the earlier years, suggests that the burden of obesity-related cancers might increase over time, given the increasing prevalence of adolescent obesity, wrote Ariel Furer, MD, of Israel Defense Forces Medical Corps, Ramat Gan, and colleagues. Their report is in The Lancet.
Obesity is a known causal factor for several types of cancer, but most studies have looked at middle-age or older individuals and had relatively short follow-up, and period effects are rarely assessed, the investigators said, noting that “the attributable burden of obesity-related cancer was previously calculated with an unverified assumption that the association remained unchanged over time.
“In contrast to this paucity of knowledge, the prevalence of youth obesity – particularly severe obesity – has increased worldwide, which parallels the rise in youth cancer incidence,” they wrote.
To address this paucity of data, the researchers reviewed medical and sociodemographic data for adolescents who were assessed at age 17 years for medical eligibility for mandatory military service, and linked that information with data from the National Cancer Registry to create a unified file. The primary study outcome was any cancer diagnosis between Jan. 1, 1967, and Dec. 31, 2012, and a secondary endpoint was all-cause mortality through Dec. 31, 2017, among those who developed cancer.
Among nearly 2.3 million participating adolescents who were evaluated for associations between body mass index at age 17 years and later cancer incidence, 1,370,020 were men with more than 29.5 million person-years of follow-up, and 928,110 were women with more than 18 million person-years of follow-up. The numbers of incident cancer cases in the men and women were 26,353 and 29,488, and the mean ages at diagnosis were 43.2 and 40.0 years, respectively, the investigators reported (Lancet. 2020 Feb 3. doi: 10.1016/S2213-8587(20)30019-X).
Adolescent obesity in men was significantly associated with midlife cancer incidence (hazard ratio, 1.26), but in women, no association was seen due to the previously reported inverse associations between obesity and cervical and breast cancers, they said.
However, when those cancers were excluded for women, the adjusted hazard ratio was similar to that for men (HR, 1.27).
Cancer incidence in both men and women increased gradually across BMI percentiles, and for both sexes, overweight BMI was associated with an increased cancer risk after 10 years of follow-up (HR, 1.14 for men, 1.22 for women after exclusion of cervical and breast cancer). Therefore, in some cases the increased cancer risk in those who were overweight as teens was evident before age 30 years, the authors noted.
Further, BMI was positively associated with greater mortality risk. For men, 5-year survival rates were 75.2% in those with adolescent BMI in the 5th-49th percentile, compared with 72.2% in those with BMI in the obesity range (95th percentile or greater), and the corresponding rates in women were 89.3% and 83.1% (HR, 1.33 and 1.89, respectively).
Of note, the investigators identified a period effect. That is, after stratification by enrollment period/cancer recording period (1967-1981/1982-1996 vs. 1982-1996/1997-2011), a stronger association was noted in individuals who entered the study during the later period, compared with those who entered in the early period (HR, 1.36 vs. 1.13; adjusted HR, 1.11 vs. 1.07 per 5 kg/m2). Possible mechanisms for this finding include environmental and nutritional factors, increased use of medical services, and changes in early cancer screening techniques, but further study is needed to verify the trend and “refine the exact nature of carcinogenic elements, compared with earlier periods,” they said.
Also of note, some cancers that were not associated with BMI in the early period, including stomach cancer, non-Hodgkin lymphoma, thyroid cancer, and colorectal and oral cavity cancers, became significantly associated with BMI in the late period.
“The projected population attributable risk percentage, using 2017 prevalence data of high BMI, was 5.1% for any cancer in men and 5.7% for cancers other than breast and cervical in women,” the researchers wrote, noting that this “is probably an underestimation, given the accentuation of the BMI-cancer association and the rapid increase in adolescent obesity prevalence within the past decade in Israel and worldwide.”
In an accompanying editorial, the journal editors noted that the findings by Dr. Furer and colleagues highlight the need to tackle obesity early in life and the need for obesity prevention strategies to reduce cancer incidence and mortality for those cancers that can be prevented by lifestyle modifications. They added, however, that care would be needed to avoid stigmatizing those with obesity, as obesity itself is a “multifactorial condition driven by social injustice and health inequalities” that most often affect those who are least able to implement lifestyle change (Lancet. 2020 Feb 3. doi: 10.106/S2213-8587(20)30031-0).
They also emphasized that the links between obesity and cancer, like those between obesity and other diseases such as diabetes, underscore the fact that noncommunicable diseases do not exist in isolation, and that tackling them requires bold action, a consolidated approach, and elimination of the environmental and social factors driving the epidemic.
The study was limited by a number of factors, including the lack of data on lifestyle factors, underrepresentation of some ethnicities, and lack of data on BMI and medical comorbidities at the time of cancer diagnosis. However, strengths of the study include the systematic data collection, narrow range of age at study entry, strict control of coexisting conditions, and high statistical power, which strengthen the generalizability of the results, the investigators said, concluding, therefore, that “[c]urrent trends of rising BMI among adolescents could constitute an important intervention target for cancer prevention.”
The authors reported having no disclosures.
SOURCE: Furer A et al. Lancet Diabetes Endocrinol. 2020 Feb 3. doi: 10.1016/S2213-8587(20)30019-X.