1 Introduction

Diabetes mellitus (DM) is a chronic condition characterized by sustained high blood glucose levels. Those with DM may be more likely than those without DM to develop other types of non-communicable diseases (NCDs) due to differences in cellular metabolism, genetic expression, hemodynamic alteration, and lifestyle factors [1, 2]. NCDs are chronic diseases that are not directly transmissible between people and can result from a combination of genetic, physiological, environmental, and behavioral factors [3, 4]. DM is a leading cause of mortality and disability globally [5]. The age-standardized DM prevalence has been increasing worldwide from 8.5% in 2011 to 9.8% in 2021, and in Indonesia from 5.1% in 2011 to 10.6% in 2021 [6]. The most prevalent microvascular complications of DM in Indonesia are neuropathy (17.6%), nephropathy (7.7%), and retinopathy (2.7%). Meanwhile, the most prevalent macrovascular complications in the country are coronary artery disease and cerebrovascular disease (both at 5.4%), followed by heart failure (5.0%) and peripheral artery disease (0.5%) [5]. Therefore, it is important to understand the associations between DM and subsequent NCDs to improve the health surveillance of those with DM.

To our knowledge, previous studies on the association between DM and other NCDs have not compared early-onset and later-onset DM diagnosed at different ages of adulthood, while examining multiple types of NCDs in a single study population. To address the literature gap on how DM diagnosed at different ages of adulthood are associated with the risks of subsequent NCDs, we used data from a large nationwide survey in Indonesia to compare the first-time diagnosis risks of several types of subsequent NCDs in those first diagnosed with DM at 20–39, 40–49, 50–59, or 60–69 years of age and those without DM history. We also determined if those with DM were diagnosed with subsequent NCDs at younger ages than those without DM. Our findings with temporality between DM and subsequent NCDs can contribute to primary care health surveillance and promote beneficial lifestyle changes in those with early-onset and later-onset DM to help prevent NCDs.

2 Materials and methods

2.1 Study design and participants

We performed a nationwide population-based analysis of 212 adults with DM history and 17,541 without DM history, using the fifth wave of the RAND Indonesian Family Life Survey (IFLS-5). The methodology and study design of IFLS-5 have been previously described [7] The IFLS-5 was conducted cross-sectionally in late 2014 and early 2015 on 16,204 households and 50,148 individuals from 13 provinces, which after survey weighting corresponded to around 83% of Indonesia’s population [7]. Because households in urban areas and provinces other than Java were oversampled, it is necessary to conduct weighted analyses to obtain estimates that represent the Indonesian population [7].

We compared the subsequent first-time NCD diagnosis risk of those first diagnosed with DM at 20–39, 40–49, 50–59, or 60–69 years of age, with those without DM history. We used age at disease diagnosis information to obtain temporality between DM and subsequent NCDs, and age at questionnaire response information to ensure participants were at risk of being diagnosed with subsequent NCDs for the first time. For example, when comparing those first diagnosed with DM at age 20–39 with those without DM history, we analyzed those 40 years of age or older at questionnaire response, and excluded those diagnosed with the NCD outcome of interest before age 40; when comparing those first diagnosed with DM at age 40–49 with those without DM history, we analyzed those 50 years of age or older at questionnaire response, and excluded those diagnosed with the NCD outcome of interest before age 50; when comparing those first diagnosed with DM at age 50–59 with those without DM history, we analyzed those 60 years of age or older at questionnaire response, and excluded those diagnosed with the NCD outcome of interest before age 60; and when comparing those first diagnosed with DM at age 60–69 with those without DM history, we analyzed those 70 years of age or older at questionnaire response, and excluded those diagnosed with the NCD outcome of interest before age 70. We defined those diagnosed with DM at 20-39 years of age as those with early-onset DM.

2.2 Ethical approval

The IFLS data are publicly available. Institutional review boards at the University of Gadjah Mada in Indonesia and the RAND Corporation in the United States have approved the IFLS study. Prior to data collection, all participants provided written informed consent [7].

2.3 Measurements of variables

The Computer-Assisted Personal Interviewing (CAPI) method was used to collect IFLS-5 data. We analyzed self-reported first-time physician diagnoses of DM and other subsequent NCDs including hypertension, lung diseases, heart diseases, arthritis, liver diseases, kidney diseases, and digestive diseases. Disease diagnoses were determined from responses of “Yes” to the question “Has a doctor/paramedic/nurse/midwife ever told you that you had [disease name]?” and “Doctor” to the question “Who diagnosed [disease name]?”. For the lung diseases variable, we included diagnoses of asthma and other non-cancerous lung conditions. The heart diseases variable was defined in the questionnaire as including heart attack, coronary heart disease, angina, or other heart problems. The kidney diseases variable was defined in the questionnaire as non-cancerous kidney diseases. The digestive diseases variable was defined in the questionnaire as non-cancerous stomach or other digestive diseases. To obtain the age at first disease diagnosis, we used the question “When was your [disease name] first diagnosed?”.

The covariates in the adjusted regression analysis included age, sex, urbanicity, and tobacco use history at questionnaire response. CAPI was used to confirm the participant’s age and sex. Urbanicity was determined by the IFLS classification of urbanicity for each province. Tobacco use history was determined by the question “Have you ever chewed tobacco, smoked a pipe, smoked cigarettes you self-rolled, or smoked cigarettes/cigars?”.

2.4 Statistical analysis

We used the weighted chi-squared test for categorical variables and weighted t-test for continuous variables in Table 1. Weighted Poisson regression was used to determine the risk ratio (RR) and 95% confidence interval (CI) for Table 2, with DM diagnosis as the independent variable and subsequent NCD diagnosis as the dependent variable. We used the weighted t-test to compare the mean age at diagnosis of subsequent NCDs in those with or without DM history for Table 3. For all weighted analyses, we utilized IFLS-5 household cross-sectional weights without attrition correction. The weighted estimates should be representative of the population at 13 Indonesian provinces in 2014 [7]. The complete case analysis was utilized to account for missing data. The multivariable weighted Poisson regression model adjusted for age at questionnaire response (continuous), sex (male/female), urbanicity (urban/rural location of household), and tobacco use history (ever/never). STATA 15.1 (StataCorp, TX) was used to conduct the analyses. GraphPad Prism 9 software was used for generating the figures. P < 0.05 was considered statistically significant.

Table 1 Characteristics of those with and without diabetes mellitus (DM) history
Full size table
Table 2 Associations between diabetes mellitus (DM) diagnosed at 20–39, 40–49, 50–59, 60–69 years of age and subsequent non-communicable diseases
Full size table
Table 3 Age at diagnosis for non-communicable diseases (NCDs) among those with or without diabetes mellitus (DM) history
Full size table

3 Results

3.1 Characteristics of the analysis populations

Table 1 shows the characteristics of the participants with or without DM diagnosis history. For those diagnosed with DM at 20–49 years of age, they were more likely to be females than those without DM history, whereas for those diagnosed with DM at 50–69 years of age, they were more likely to be males than those without DM history. Those with DM were also more likely to live in urban areas and have history of tobacco use than those without DM history. The most common subsequent NCDs for those diagnosed with DM at 20–39 years of age were hypertension (34.7%), digestive diseases (22.6%), and arthritis (20.1%). The most common subsequent NCDs for those diagnosed with DM at 40–49 years of age were hypertension (42.0%), digestive diseases (22.2%), and heart diseases (11.8%). For those diagnosed with DM at 50–59 years of age, the most common subsequent NCDs were hypertension (47.1%), heart diseases (17.3%), and arthritis (13.1%). For those diagnosed with DM at 60–69 years of age, the most common subsequent NCDs were hypertension (58.7%), lung diseases (15.0%), and digestive diseases (11.0%).

3.2 Associations between DM and NCDs

Table 2 presents the crude and adjusted associations between DM diagnosed at 20–39, 40–49, 50–59, or 60–69 years of age and subsequent NCDs. The crude and adjusted estimates were consistent with each other. Those with DM generally had higher risks of subsequent NCDs compared with those without DM history. Those diagnosed with DM in all age groups had significantly higher risk of hypertension compared with those without DM history. Those diagnosed with DM at 20–39 years of age had especially higher risks of lung diseases and arthritis compared with those without DM history, and those diagnosed with DM at 20–49 years of age had especially higher risks of digestive diseases compared with those without DM history. On the other hand, those diagnosed with DM at 40 years of age or above had especially higher risks of liver and kidney diseases compared with those without DM history. Those diagnosed with DM at 40–59 years of age had especially higher risk of heart diseases compared with those without DM history. Figure 1 summarizes the adjusted RR in Table 2 with line graphs.

Fig. 1
figure 1

Adjusted associations between diabetes mellitus (DM) diagnosed at 20–39, 40–49, 50–59, and 60–69 years of age and subsequent non-communicable diseases. Analyses were performed using weighted Poisson regression. The adjusted model controlled for age (continuous), sex (male/female), urbanicity (rural/urban), and tobacco use (ever/never). *P < 0.05, **P < 0.01, ***P < 0.001

Full size image

3.3 Mean age at NCD diagnosis

Table 3 and Fig. 2 compare the mean age at diagnosis for the subsequent NCDs in those with or without DM history. Overall, those with DM were diagnosed with subsequent NCDs at younger ages compared with those without DM history, and most of these differences in the mean age at diagnosis reached statistical significance. The difference in the mean age at NCD diagnosis between those diagnosed with DM at 20–39 years of age and those without DM history ranged from 3.2 to 8.3 years. The difference in the mean age at NCD diagnosis between those diagnosed with DM at 40–49 years of age and those without DM history ranged from 4.5 to 9.5 years. The difference in the mean age at NCD diagnosis between those diagnosed with DM at 50–59 years of age and those without DM history ranged from 3.3 to 9.2 years. Finally, the difference in the mean age at NCD diagnosis between those diagnosed with DM at 60–69 years of age and those without DM history ranged from 2.5 to 6.5 years.

Fig. 2
figure 2

Age at diagnosis for non-communicable diseases (NCDs) among those with or without diabetes mellitus (DM) history. Analyses were performed using weighted t-test. *P < 0.05, **P < 0.01, ***P < 0.001

Full size image

4 Discussion

In this nationwide study of the Indonesian population on the associations between DM and subsequent NCDs, we were able to compare how DM first diagnosed at different ages of adulthood were associated with subsequent first-time NCD diagnoses. We also found those with DM were diagnosed with subsequent NCDs at younger ages compared with those without DM history. These findings contribute to the health surveillance of those with early-onset and later-onset DM diagnosed at various stages of adulthood, which may help prevent the subsequent development of other NCDs.

4.1 Hypertension risk

In our study, participants with DM diagnosed in all age groups had significantly higher risk of subsequent hypertension compared with those without DM history, which suggests vigilant hypertension surveillance for both early-onset and later-onset DM patients. Previous studies have found hypertension and DM influenced each other’s development [8], along with a study from the UK that found hypertension affects around 70% of diabetes patients and is roughly twice as common in those with DM as those without DM [9]. Around 40% of type 2 diabetes patients are hypertensive by the age of 45, rising to 60% by the age of 75 [9]. The association between DM and higher risk of hypertension may be explained by vascular remodelling and dysfunction caused by the exaggerated oxidative stress, inflammation, and fibrosis in patients with DM [10].

4.2 Heart disease risk

We found those with later-onset DM diagnosed at 40–49 and 50–59 years of age had significantly higher risk of subsequent heart diseases compared with those without DM history. Although the association between DM and heart diseases is well-known, few studies have focused on DM diagnosed at various ages of adulthood and compared their heart disease risk with those without DM history. A study from Taiwan found that type 2 DM patients diagnosed at below 60 years of age had greater mortality, as well as more macrovascular and microvascular complications than patients diagnosed at 60 years of age or above [11]. We also found the mean age at diagnosis of subsequent heart diseases was lower in patients with DM regardless of their age at DM diagnosis. Another study showed those with DM had younger mean ages of acute myocardial infarction, stroke, or death from any cause, compared with those without DM [12]. Our findings support more vigilant surveillance of heart diseases in adults diagnosed with DM before 60 years of age.

4.3 Liver and kidney disease risk

We also found those with later-onset DM diagnosed at 40 years of age or above had especially higher risk of subsequent liver or kidney diseases compared with those without DM history. Our findings for liver diseases were consistent with studies from other Asian countries that found liver disease prevalence was higher among type 2 DM patients, particularly in younger and middle-aged adults [13, 14]. However, those studies did not categorize by the age at DM diagnosis of the participants, and some did not establish temporality [13, 14]. For kidney diseases, a South Korean study of elderly individuals showed chronic kidney disease was common in diagnosed with type 2 DM [15]. However, an Australian study indicated younger-onset type 2 DM increased the risk of end-stage kidney disease due to longer DM duration [16]. A study from Singapore found 53% of type 2 DM patients had kidney diseases [17] and a study from the Netherlands found renal microvascular dysfunction rate was nearly four times higher in individuals with type 2 DM than those without DM [18].

4.4 Lung disease risk

In our study, participants diagnosed with early-onset DM at 20–39 years of age had significantly higher risk of subsequent lung diseases compared with those without DM history. The association between DM and asthma has been previously established [19]. Other lung diseases such as pulmonary hypertension are associated with DM due to microvascular and macrovascular damage that could affect the pulmonary vasculature [19]. DM is also associated with lung diseases such as chronic obstructive pulmonary disease and idiopathic pulmonary fibrosis [19]. Further studies are needed to verify our findings on the association between early-onset DM and lung diseases.

4.5 Arthritis risk

We also found participants diagnosed with early-onset DM at 20–39 years of age had significantly higher risk of subsequent arthritis compared with those without DM history. A meta-analysis found an association between rheumatoid arthritis and higher risk of DM [20]. There is also increasing evidence from the clinical and laboratory settings that suggests a negative influence of DM on the development and severity of osteoarthritis [21]. A systematic review found the risk of osteoarthritis was higher in those with DM than those without DM, and the risk of DM was higher in those with osteoarthritis than those without osteoarthritis [22]. Future studies with information on specific types of arthritis are needed to confirm our findings and better understand the directionality of the association between DM and arthritis.

4.6 Digestive diseases risk

We found adults with DM diagnosed before 50 years of age had significantly higher risk of subsequent digestive diseases compared with those without DM history. A population-based study from Australia found that gastrointestinal symptoms were more common in those with both types of DM than in those without DM [23]. Although cross-sectional epidemiological studies have revealed that gastrointestinal symptoms were linked with poor glycemic management, they did not assess for temporality [23, 24]. Several mechanisms may explain these findings. DM-related gastrointestinal tract manifestations are typically caused by unusual gastrointestinal motility from diabetic autonomic neuropathy affecting the gastrointestinal tract [25]. Anti-hyperglycemic agents such as metformin may also play a role in the gastrointestinal symptoms of DM patients [26].

4.7 Strengths and limitations

Our analysis of a nationwide population-based dataset with age at disease diagnosis information allowed us to obtain temporality between the diagnosis of DM and subsequent NCDs, and to compare the risk of NCDs in those with DM diagnosed at different ages of adulthood with those without DM history. However, this study has several limitations. Although we only included physician diagnoses to reduce disease misclassification, these self-reported diagnoses did not include undiagnosed individuals and diagnosed individuals who could not participate in the study. The Indonesian Ministry of Health reported that the DM prevalence was 6.9% in 2013 based on blood glucose testing, whereas the self-reported DM prevalence was only 2.1% [27]. Another limitation is that we lacked information on the types, duration, and severity of DM. Although we could not distinguish between type 1 and type 2 DM, our study results are likely mostly generalizable to type 2 DM patients, because the proportion of type 2 DM out of all DM types is above 90% [28, 29]. We also lacked detailed information on the specific types of other NCDs, because the survey only asked general questions on the NCD types. Lastly, even though the results of this analysis will be useful for health surveillance, we could not fully understand the causal associations between DM and NCDs due to unmeasured confounding.

5 Conclusion

The results from this study contribute to the understanding of the subsequent NCD risks in those diagnosed with DM at different stages of adulthood, which may improve primary care health surveillance and promote beneficial lifestyle changes in those with DM to help prevent NCDs. Hypertension was the most common subsequent NCD among both individuals with DM and those without DM history. However, the risk of subsequent hypertension was significantly higher among those with DM. Furthermore, individuals with early-onset DM diagnosed at 20–39 years of age exhibited significantly elevated risks of arthritis, digestive diseases, and lung diseases. Additional research using longitudinal data from a large cohort study is required to better understand the associations between DM diagnosed at various ages and the risks of subsequent NCDs. Further studies are also needed to examine these associations by DM type and severity, and to better understand the mechanisms by which DM at different ages increases the risks of different NCDs.