**3. Results**

#### *3.1. Baseline Characteristics of Study Cohort*

The study included 840 subjects (54.5% female), whose baseline characteristics are shown in Table 1 and in Tables S1 and S2. The median follow-up time for mortality was 6.9 years (interquartile range 4.6–8.5 years). The average age of the cohort was 76.1 ± 6.8 years, with no significant difference between men and women (*p* = 0.39). Men, on average, had higher serum IGF-1 levels (*p* < 0.001) and IGF-1/IGFBP-3 molar ratios (*p* < 0.001), but lower IGFBP-1 (*p* < 0.001) and IGFBP-3 (*p* < 0.001) compared to women. The average age was significantly older in subjects with low IGF-1 compared to those with high IGF-1 in the combined cohort (*p* = 0.04) and among men in the sex-stratified analysis (*p* = 0.02), (Table S2). Over the course of follow-up, 13.9% of study participants died. At baseline, prevalence of morbidities was as follows: cardiovascular disease 12.7%, diabetes 10.5%, cancer 22.1%, and MDCI 3.0% (Table S1).


**Table 1.** Baseline characteristics of study cohort. All *p*-values are for comparisons between males and females.

#### *3.2. IGF-Associated Proteins and Mortality: Low IGFBP-3 and High IGF-1*/*IGFBP-3 Molar Ratio Predict Mortality Risk*

In unadjusted analysis, baseline IGF-1 levels were not predictive of mortality risk (Figure 1a–c). Upon adjustment for baseline age, we observed a non-significant trend towards higher mortality hazard with high IGF-1 levels in women (HR = 1.28, 95% CI 0.96–1.71, *p* = 0.09; Figure 2c). High levels of IGFBP-1, compared to low levels, were associated with significantly higher mortality risk in the overall cohort (*p* < 0.001) and among men (*p* < 0.001), (Figure 1d–f), but the associations became non-significant upon adjustment for age and sex (Figure 2a–c). On the other hand, high levels of IGFBP-3 predicted a lower mortality risk in an unadjusted analysis of the overall cohort (*p* < 0.001), as well as in men (*p* = 0.005) and women (*p* = 0.003) (Figure 1g–i). The difference in mortality risk between subjects with high vs. low IGFBP-3 remained significant upon adjusting for age and sex in the overall cohort (HR 0.82, 95% CI 0.680–0.998, *p* = 0.048), while in sex-stratified analysis associations retained the same direction but lost statistical significance (Figure 2a–c). Further adjustment for IGF-1 did not significantly alter the association between IGFBP-3 and mortality in the overall cohort (HR 0.71, 95% CI 0.56–0.89, *p* = 0.003) and strengthened it in women (HR 0.60, 95% CI 0.43–0.84, *p* = 0.003), but the association in men remained non-significant (HR 0.85, 95% CI 0.62–1.17 *p* = 0.32). High IGF-1/IGFBP-3 molar ratio, which is an estimate of circulating free IGF-1, was associated with higher mortality risk in the overall cohort (*p* = 0.002) and in women (*p* = 0.003) in unadjusted analysis (Figure 1j–l), and these associations persisted upon adjustment for age and sex (HR 1.28, 95% CI 1.05–1.57, *p* = 0.02 and HR 1.53, 95% CI 1.12–2.09, *p* = 0.007, respectively) (Figure 2a–c). After exclusion of 5 participants (4 males, 1 female) who died during the first year of follow-up, associations between IGF-1 and related proteins with mortality remained largely unchanged (data not shown).

Stratification by median participant baseline age showed consistent associations of IGF-related proteins and mortality between the two age groups (Figure S1), with the exception of IGFBP-1, which was significantly positively associated with mortality in younger (HR 1.60, 95% CI 1.04–2.46, *p* = 0.03) but not in older participants (HR 0.98, 95% CI 0.78–1.22, *p* = 0.85). High IGF-1, on the other hand, was more strongly associated with mortality among older women (HR 1.47, 95% CI 1.03–2.08, *p* = 0.03) compared with younger women (HR 1.10, 95% CI 0.64–1.87, *p* = 0.73).

**Figure 1.** Insulin-like growth factor (IGF)-associated proteins and mortality. Unadjusted survival curves for individuals with high and low levels of IGF-1 (**<sup>a</sup>**. combined cohort; **b**. males; **c**. females), IGFBP-1 (**d**–**f**), IGFBP-3 (**g**–**i**), and IGF-1/IGFBP-3 molar ratio (**j**–**l**).

**Figure 2.** IGF-associated proteins and mortality hazard. Sex and age-adjusted survival hazards for combined cohort (**a**) and age-adjusted survival hazard for males (**b**) and females (**c**) with high levels of IGF-associated proteins as compared to individuals with low levels.

#### *3.3. IGF-Associated Proteins and Morbidity: High IGF-1 Predicts Risk for MDCI and Age-Related Composite Morbidity while Low IGFBP-1 Predicts Risk for Diabetes*

High IGF-1 levels, compared to low IGF-1 levels, were associated with greater risk for incident MDCI in the overall cohort (*p* = 0.04) and in men (*p* = 0.045). These associations remained significant after adjusting for baseline age and sex, with HR 1.56, 95% CI 1.08–2.26, *p* = 0.02 and HR 1.81, 95% CI 1.04–3.16, *p* = 0.04 for MDCI in the overall cohort and in men, respectively. Similarly, we observed a greater risk with high IGF-1 for composite incident morbidity in the overall cohort (*p* = 0.04) and in men (*p* = 0.03) (Figure 3), which remained significant after adjustments (HR 1.242, 95% CI 1.004–1.538, *p* = 0.046 and HR 1.44, 95% CI 1.04–2.01, *p* = 0.03, respectively; Figure 4). High IGFBP-1, compared to low IGFBP-1 level, was also associated with higher risk for incident MDCI in men (*p* = 0.004), but not in the overall cohort (*p* = 0.11) or in women (*p* = 0.43), (Figure 5). After adjusting for age and sex, the association between IGFBP-1 level and MDCI hazard became non-significant (Figure 6). On the other hand, high IGFBP-1 was associated with reduced diabetes risk in the overall cohort in unadjusted analysis (*p* = 0.01), (Figure 5). In age- and sex-adjusted analysis, high IGFBP-1 remained significantly associated with protection from incident diabetes in the overall cohort (HR 0.50, 95% CI 0.29–0.88, *p* = 0.02) and in men (HR 0.31, 95% CI 0.10–0.92, *p* = 0.03; Figure 6). These associations persisted upon inclusion of IGF-1 as a covariate in the age- and sex-adjusted model (HR 0.50, 95% CI 0.29–0.89, *p* = 0.02 in the overall cohort; HR 0.30, 95% CI 0.10–0.89, *p* = 0.03 in men). However, when body mass index (BMI) and insulin levels were added to the model, the association between IGFBP-1 and diabetes was attenuated and no longer significant (HR 0.47, 95% CI 0.21–1.03, *p* = 0.06 in the overall cohort; HR 0.66, 95% CI 0.26–1.67, *p* = 0.38 in men). Levels of IGFBP-3 were not significantly associated with risk of any of the investigated age-related diseases (Figures S2 and S3). Associations between high IGF-1/IGFBP-3 ratio and MDCI risk were in the same directions as those between IGF-1 and MDCI, but they reached statistical significance only among women in both unadjusted (*p* = 0.04) and age-adjusted analyses (HR 1.81, 95% CI 1.03–3.21, *p* = 0.04), (Figures S4 and S5).

**Figure 3.** IGF-1 levels and morbidity. Unadjusted survival curves for multiple-domain cognitive impairment (MDCI) (**<sup>a</sup>**. combined cohort; **b**. males; **c**. females), diabetes (**d**–**f**), cardiovascular disease (**g**–**i**), cancer (**j**–**l**), and composite incident morbidity (**<sup>m</sup>**–**<sup>o</sup>**) in individuals with high and low levels of IGF-1.

**Figure 4.** IGF-1 and morbidity hazard. Sex and age-adjusted morbidity hazards for all individuals in cohort (**a**), and age-adjusted morbidity hazards for males (**b**) and females (**c**) with high levels of IGF-1 as compared to individuals with low levels.

**Figure 5.** IGFBP-1 levels and morbidity. Unadjusted survival curves for MDCI (**<sup>a</sup>**. combined cohort; **b**. males; **c**. females), diabetes (**d**–**f**), cardiovascular disease (**g**–**i**), cancer (**j**–**l**), and composite incident morbidity (**<sup>m</sup>**–**<sup>o</sup>**) in individuals with high and low levels of IGFBP-1.

**Figure 6.** IGFBP-1 and morbidity hazard. Sex and age-adjusted morbidity hazards for all individuals (**a**), and age-adjusted morbidity hazard for males (**b**) and females (**c**) with high levels of IGFBP-1 as compared to individuals with low levels.
