Additive interaction of family medical history of diabetes with hypertension on the diagnosis of diabetes among older adults in India: longitudinal ageing study in India - BMC Public Health

Table 1 presents the sample characteristics and percentage distribution of diabetes and its treatment among individuals aged 45 and above. A proportion of 34.43% of the participants were 65 years and above. Approximately 54% of the sample population was female. About 50.45% of the sample had no education during the survey. A large proportion of the sample (73.93%) were in marital union during the survey. Further nearly 70% of the participants were living in rural areas.

Table 1 Sample characteristics and percentage distribution of diabetes and its treatment by background characteristics among individuals aged 45 years and above, Longitudinal Ageing Study in India (LASI, 2017-18) Full size table

Table 1 also depicts the prevalence of diabetes among adults aged 45 years and above. The overall prevalence of diabetes was 11.9% (95% CI: 11.2, 12.7) and 82.5% of the individuals with diabetes were taking treatment. The prevalence of diabetes (16.0%) was higher among individuals in 65–74 years of age group. The prevalence of diabetes was higher among individuals with higher education (19.3%) than no education (7.8%). Additionally, diabetes was more prevalent among individuals living in urban areas (20.7%) than those in rural areas (8.2%). The results show that the prevalence of diabetes was higher among physically inactive (13.5% vs. 11.1%), overweight/obese participants (21.5% vs. 10.3%) and individuals with high cholesterol (34.6% vs. 11.4%) than their counterparts.

Moreover, diabetes was more prevalent among individuals with self-reported hypertension (27.5% vs. 6.2%) than their counterparts. The prevalence of hypertension was higher among hypertensive individuals who were taking treatment (32.7% vs. 14.2%) compared with individuals with self-reported hypertension who were not taking treatment.

Furthermore, we found that the prevalence of diabetes was 3 times higher among individuals with family history of diabetes (27.8% vs. 9.2%) than those without family history. Similarly, parental and sibling history of diabetes had 3 times higher prevalence of diabetes (31.1% and 31.8%) compared with those without parental and sibling history of diabetes (9.9% and 10.4%). We observed that the prevalence of diabetes was higher among individuals with father (36.5% vs. 10.8%), mother (29.0% vs. 10.9%), brother (34.6% vs. 10.7%) and sister (36.4% vs. 11.2%) medical history of diabetes compared with those without medical history of diabetes.

Figure 2 illustrate that the prevalence of diabetes was higher among female participants with family history of diabetes, including parental, sibling, father, mother, brother, sister medical history of diabetes than males. Hypertension was more prevalent among male participants than females. On the other hand, Fig. 3 shows that the coexistence of family medical history and hypertension was higher among male participants than females.

Fig. 2 The prevalence of diabetes by key predictor variables and gender Full size image

Fig. 3 The prevalence of diabetes by joint effect of key predictor variables and gender HTM, hypertensive individuals who were taking medication; No-coexistence, either hypertension or family history of diabetes; FH, family history Full size image

Table 2 represents the prevalence of diabetes by key predictor variables among individuals aged 45 and above.

Table 2 The prevalence of diabetes by key predictor variables among individuals aged 45 and above Full size table

The prevalence of diabetes based on the presence or absence of family medical history of diabetes and hypertension

Self-reported hypertension and family medical history of diabetes

The result presents that the prevalence of diabetes was significantly higher among individuals with hypertension and family history of diabetes (46.6%, 95% CI: 39.7–53.6) than those without the coexistence of family history of diabetes and hypertension (9.9%, 95% CI: 9.5–10.4), individuals with hypertension and without a family history of diabetes (22.7%, 95% CI: 21.2–24.2), individuals with family history of diabetes and without hypertension (16.5%, 95% CI: 14.5–18.7), and individuals without the presence of both family history of diabetes and hypertension (4.7%, 95% CI: 4.3–5.1).

Hypertensive individuals (taking antihypertensive medication) and family medical history of diabetes

The result presents that the prevalence of diabetes was significantly higher among hypertensive individuals (those who were taking medication for hypertension) with family history of diabetes (53.1%, 95% CI: 45.2–60.9) than those without the coexistence of family history of diabetes and hypertension (10.1%, 95% CI: 9.6–10.6), hypertensive individuals and without a family history of diabetes (26.9%, 95% CI: 25.0–28.8), individuals with family history of diabetes and without hypertension (17.0%, 95% CI: 15.2–19.1), and individuals without the presence of both family history of diabetes and hypertension (5.5%, 95% CI: 5.1–5.9).

The prevalence of diabetes (who were taking medication) based on the presence or absence of family medical history and hypertension

Self-reported hypertension and family medical history of diabetes

The result presents that the prevalence of diabetes (those who were also taking medication) was significantly higher among individuals with hypertension with family history of diabetes (41.9%, 95% CI: 34.6–49.5) than those without the coexistence of family history of diabetes and hypertension (8.0%, 95% CI: 7.6–8.5), individuals with hypertension and without a family history of diabetes (18.1%, 95% CI: 16.7–19.6), individuals with family history of diabetes and without hypertension (13.5%, 95% CI: 11.8–15.4), and individuals without the presence of both family history of diabetes and hypertension (3.8%, 95% CI: 3.5–4.2).

Hypertensive individuals (taking antihypertensive medication) and family medical history of diabetes

The result presents that the prevalence of diabetes (those who were also taking medication) was significantly higher among hypertensive individuals (those who were taking medication for hypertension) with family history of diabetes (50.4%, 95% CI: 42.2–58.5) than those without the coexistence of family history of diabetes and hypertension (8.0%, 95% CI: 7.6–8.5), hypertensive individuals and without a family history of diabetes (23.7%, 95% CI: 21.8–25.7), individuals with family history of diabetes and without hypertension (13.0%, 95% CI: 11.5–14.7), and individuals without the presence of both family history of diabetes and hypertension (3.9%, 95% CI: 3.6–4.3).

Table 3 presents the unadjusted and adjusted logistic regression estimates for diabetes by family medical history of diabetes (overall, parental history, father, mother, sibling history, brother and sister) and hypertension among individuals aged 45 years and above. In the adjusted model, the results indicate that individuals with hypertension (AOR: 3.95, CI: 3.52–4.44) had 3.95 times higher odds of having diabetes than those without hypertension.

Table 3 Logistic regression estimates for self-reported diabetes and individuals with medication by family history and hypertension Full size table

Our findings show that individuals with family history of diabetes (AOR: 2.47, CI: 2.11 2.89) had significantly 2.47 times higher odds of having diabetes than those without family history. Similarly, individuals with parental and sibling medical history had 2.79 and 2.22 times higher odds of having diabetes, respectively, than those without parental and sibling history. Moreover, our findings further demonstrate that individuals with father, mother, brother, sister medical history of diabetes had significantly 3.29, 2.16, 2.37, and 2.40 times higher odds of having diabetes, respectively compared with those without family history.

Table 4 provides the multivariable logistic regression estimates for diabetes by the joint effect of family history of diabetes and hypertension. The table also provides unadjusted and adjusted models of additive interaction of family history of diabetes with hypertension on diabetes. The current study provides estimates of the interaction effect on the additive scale for all models of hypertension (self-reported hypertension, and hypertensive individuals receiving treatments) in two different samples (self-reported diabetes and individuals with diabetes receiving medication).

Table 4 Additive interaction between family history of diabetes and hypertension on the diagnosis of diabetes among individuals aged 45 and above Full size table

In the additive model, the interaction effects between family history of diabetes and hypertension were found to be significantly positive, which demonstrates that the combined effect of two exposures (family history of diabetes and hypertension) is larger than the sum of the individual effects on the diagnosis of diabetes.

Interaction effect between family medical history and hypertension on diabetes

Our study findings show that the adjusted odds ratio (AOR) of the joint effect between family medical history of diabetes and hypertension on the diagnosis of diabetes was 9.28 (95% CI: 7.51–11.46). In the adjusted model, the relative excess risk due to interaction (RERI) was 3.5 (95% CI: 1.52–5.47), which indicates that there is a significant positive interaction between family history and hypertension on the diagnosis of diabetes. The attributable proportion due to interaction (AP) value was 37% (0.37; 95% CI: 0.22–0.51), which suggests that a significant proportion of individuals with diabetes in the population can be attributed to the interaction between family medical history of diabetes and hypertension. The synergistic effect index (S) was 1.69 (95% CI: 1.31–2.18), further supporting a significant synergistic effect.

Furthermore, our findings further support the interaction effect on all the three measures based on hypertensive individuals who were taking medication. In the adjusted model, the RERI, AP and S values for diabetes were 3.79 (95% CI: 1.58-6.01), 40% (0.40; 95% CI: 0.25–0.55), and 1.81 (95% CI: 1.37–2.4) respectively, which indicates that there is a significant positive interaction between family history of diabetes and hypertension (taking antihypertensive medication) on the diagnosis of diabetes.

All the three measures of interaction, the RERI, AP and S, show significant positive interaction on the additive scale demonstrating consistent results in all the models (self-reported hypertension, and hypertensive individuals receiving medication) even after adjustment with potential confounding factors.

Interaction effect between family medical history and hypertension on the diagnosis of diabetes (taking medication for diabetes)

The interaction effect of family medical history of diabetes and hypertension is further supported by selecting the individuals with diabetes who were taking medications.

Interaction effect between family medical history and hypertension on the diagnosis of diabetes

The results show that the adjusted odds ratio (AOR) of the joint effect between family medical history of diabetes and hypertension on the diagnosis of diabetes was 8.5 (95% CI: 6.82–10.61). In the adjusted model, the RERI, AP, and S for diabetes were 3.24 (95% CI: 1.46–5.03), 38% (0.38; 95% CI: 0.24–0.52), and 1.76 (95% CI: 1.34–2.3) respectively, which indicates that there is a significant positive interaction between family history of diabetes and hypertension on the diagnosis of diabetes.

Interaction effect between family medical history and hypertensive individuals taking antihypertensive medication on the diagnosis of diabetes

Furthermore, our findings further support the interaction effect on all the three measures based on hypertensive individuals who were taking medication. In the adjusted model, the RERI, AP and S values for diabetes (with medication) were 4.91 (95% CI: 2.33–7.49), 44% (0.44; 95% CI: 0.30–0.58), and 1.94 (95% CI: 1.47–2.56) respectively, which indicates that there is a significant positive interaction between family history of diabetes and hypertension on the diagnosis of diabetes. The findings present that the combined effect of family medical history of diabetes and hypertension on the risk of developing diabetes is greater than the sum of their individual effects.

All the three measures of interaction, the RERI, AP and S, show significant positive interaction on the additive scale demonstrating consistent results in all the models (self-reported hypertension, hypertensive individuals receiving medication) even after adjustment with potential confounding factors. Consequently, indicates that the combined effect of family history of diabetes and hypertension is more than the sum of the individual effects on the risk of developing diabetes among older adults aged 45 years and above in India.