• 2018-07
  • 2020-07
  • 2020-08
  • br Results br The baseline characteristics


    3. Results
    The baseline characteristics of participants included in the final analysis are described in Table 1. A larger portion of women used any sedative-hypnotic medication compared with men; among women, 53% were prescribed any one of the sedative-hypnotics
    Table 1
    General characteristics of 236,759 participants included in the analysis.
    Non-usersa By dose
    By hypnotic groups
    Any hypnotic Low dose Medium High dose
    Use of GABAAc GABAAc hypnotic Non-GABA
    or non-GABA groupd use 30 DDD hypnoticd
    Insurance premium
    DDD, defined daily dose; GABA, gamma-aminobutyric acid; GABAA, GABA ASP-1517 agonist; SD, standard deviation. a Non-users: <30 DDD of any sedative-hypnotic medication.
    b Any hypnotic use: e30 DDD of any sedative-hypnotic medication.
    c GABAA group: GABA agonist group including benzodiazepine and zolpidem.
    d Non-GABA group: Including antidepressants and antipsychotics.
    included in the analysis, whereas 47% of men were prescribed any sedative-hypnotic. Compared with non-users, more participants with hypnotic use were in the low-insurance-premium group and participants in the low-insurance-premium group were prescribed relatively higher doses of sedative-hypnotics. In addition, partici-pants with higher doses of sedative-hypnotic use showed increased numbers of comorbidities. Both men and women showed a higher frequency of prescription of GABAA group medications compared with non-GABA group medications; the difference ranged from 10-times higher among men to 16.5-times higher among women.
    Overall cancer incidence and HRs in relation to sedative-hypnotic use are presented in Table 2. Participants who were
    Table 2
    Sedative-hypnotic medication use and overall incidence of cancer. 
    exposed to sedative-hypnotics had an 8% higher incidence of overall cancer after adjusting for sex, comorbidities, and socio-economic status (results not shown). When stratified by sex, only women had a statistically significant increase in risk for new-onset cancer before and after full adjustment (adjusted HR ¼ 1.21, 95% CI ¼ 1.13e1.29). When considering types of sedative-hypnotics, both men and women in the GABAA group (HR ¼ 1.38, 95% CI ¼ 1.16e1.64 in men; HR ¼ 1.60, 95% CI ¼ 1.27e2.03 in women) had statistically significant higher HRs for overall cancer incidence. Among women, exposure to sedative-hypnotics with a dosage in the 180- to 360-DDD range increased the incidence for overall cancer by a statistically significant 30%. However, there was not a
    Person-year Case (N) Incidence rate (%) Crude HR
    Adjusted HR*
    Sedative-hypnotic use
    Sedative-hypnotic use
    No hypnotic usea
    b 30 DDD of any sedative-hypnotic medication.
    c GABAA group: GABA agonist group including benzodiazepine and zolpidem.
    d Non-GABA group: Including antidepressants and antipsychotics.
    statistically significant doseeresponse pattern for overall cancer incidence as sedative-hypnotic dosage increased.
    In subgroup analyses that restricted the exposure to the GABAA group only, a statistically significant, elevated HR for overall cancer was found for women (results not shown). In addition, we observed a statistically significant, increased HR for participants exposed to sedative-hypnotics in the 180- to 360-DDD range group (HR ¼ 2.24, 95% CI ¼ 1.60e3.14).
    4. Discussion
    We examined the relationship between exposure to sedative-hypnotic medications and risk of cancer using data from 236,759 participants selected from Korea's nationwide health insurance database. We observed an increased risk for overall cancer; how-ever, we found no doseeresponse relationship with the dosage of the sedative-hypnotic medications under study. Participants exposed to GABAA sedative-hypnotics showed a larger increase in HR compared with non-GABA sedative-hypnotics. In this study, women with any sedative-hypnotic use had a significantly increased risk for thyroid, breast, ovarian and lung cancer, and marginally decreased risk for cervical cancer. Among men with any sedative-hypnotic use, we observed increased risk for prostate, brain and lung cancer, and decreased risk of stomach cancer.
    However, our results should be interpreted with caution. Although we adjusted the comorbidity status using the Charlson Comorbidity Index in the final model, it is not clear whether the sedative-hypnotic medication or conditions related to poor sleep affects cancer incidence.
    Table 3
    Sedative-hypnotic use and incidence of individual cancers. 
    Early caseecontrol studies [4e6,18e20] that evaluated the as-sociation between sedative-hypnotic use and certain cancers observed no significant results. These studies were conducted with relatively small sample sizes, and thus lacked sufficient power to observe significant associations. More recent studies [7e10,21] that were conducted with a prospective or retrospective cohort design in large populations have reported several significant results be-tween sedative-hypnotic use and cancer incidence. Recent meta-analysis including four cohort studies and 18 caseecontrol studies found statistically significant increased association between benzodiazepine use and cancer incidence [22]. For specific cancer sites, one cohort study [21] and one large caseecontrol study [9] found significant associations between benzodiazepine use and increased prostate cancer risk. In our study, we also found a significantly increased risk for prostate cancer in the GABAA group, which included benzodiazepine and zolpidem, but not in the non-GABA group. In our sample, both men and women showed increased incidence of lung cancer; two large caseecontrol studies [7,9] reported increased lung cancer risk in their benzodiazepine-prescribed group. We could not find an increased risk of liver cancer among both men and women who were prescribed sedative-hypnotics, which is inconsistent with results from Kao et al. [21], and two other previous studies [7,9] that reported increased risk for liver cancer among both men and women sedative-hypnotic users. Contrary to our results, female cancers, such as breast and ovarian cancer, have shown no significant results in several other studies [4,5,20,23,24]. However, three studies [25e27] did show an increased risk for breast and ovarian cancer in benzodiazepine-prescribed groups. Cervical cancer showed a marginally decreased risk in one study [9], which is consistent with our findings; however, a recent study conducted in large Taiwanese population reported the opposite results for cervical cancer [21]. Thyroid cancer was assessed in only one other study [25], which reported non-significant results for sedative-hypnotic users, whereas our study found an increased risk for thyroid cancer among women.