Baseline characteristics
After extracting 101, 316 participants from the NHANES 1999-2018 cohorts, we applied inclusion and exclusion criteria, and 16, 670 participants were found to be eligible. A thorough flow diagram was shown in Figure 1. The participants were divided into quartiles based on their ALI levels. The average ALI values for Quantile1 group, Quantile2 group, Quantile3 group, and Quantile4 group were 34.16, 53.60, 72.29, and 112.75, respectively. The average age of these 16, 670 participants was 55.73 years and a mean ALI value of 69.93.
As ALI increased, age, neutrophil count, and neutrophil-to-lymphocyte ratio (NLR) decreased, while body mass index (BMI) and lymphocyte count increased. Higher ALI quartiles showed a greater proportion of females, Non-Hispanic Blacks, and never-smokers. Notably, cardiovascular disease prevalence, all-cause mortality, and cardiovascular disease mortality rates significantly decreased with increasing ALI. Detailed results were displayed in Table 1.
Table 1. Baseline characteristics of the study population by ALI quartiles
Characteristic
|
Overall
|
Q1
|
Q2
|
Q3
|
Q4
|
P-value
|
N
|
16670
|
3792
|
4001
|
4181
|
4696
|
|
ALI
|
69.93 ± 34.75
|
34.16 ± 7.53
|
53.60 ± 4.90
|
72.29 ± 6.34
|
112.75 ± 25.19
|
<0.001
|
Age (years)
|
55.73 ± 16.55
|
61.65 ± 16.85
|
56.23 ± 16.55
|
53.80 ± 16.05
|
52.26 ± 15.37
|
<0.001
|
BMI (kg/m²)
|
30.14 ± 6.85
|
26.89 ± 5.25
|
29.19 ± 5.86
|
30.74 ± 6.53
|
33.04 ± 7.68
|
<0.001
|
ALT (U/L)
|
26.20 ± 28.15
|
23.56 ± 47.75
|
25.65 ± 18.43
|
26.86 ± 19.25
|
28.19 ± 18.60
|
<0.001
|
AST (U/L)
|
25.88 ± 21.15
|
25.47 ± 35.88
|
25.43 ± 14.48
|
25.82 ± 14.28
|
26.64 ± 13.70
|
0.024
|
BUN (mmol/L)
|
5.16 ± 2.15
|
5.64 ± 2.67
|
5.15 ± 2.10
|
5.05 ± 1.93
|
4.87 ± 1.81
|
<0.001
|
Creatinine (μmol/L)
|
81.05 ± 34.37
|
87.09 ± 50.77
|
79.74 ± 28.23
|
79.28 ± 26.81
|
78.85 ± 27.47
|
<0.001
|
eGFR
|
89.73 ± 22.08
|
84.02 ± 24.06
|
90.26 ± 21.74
|
91.67 ± 21.40
|
92.17 ± 20.42
|
<0.001
|
Neutrophils (10³ cells/μL)
|
4.21 ± 1.60
|
5.35 ± 1.77
|
4.49 ± 1.39
|
3.97 ± 1.28
|
3.26 ± 1.15
|
<0.001
|
Lymphocytes (10³ cells/μL)
|
2.13 ± 0.71
|
1.62 ± 0.50
|
1.98 ± 0.54
|
2.22 ± 0.60
|
2.58 ± 0.76
|
<0.001
|
NLR
|
2.14 ± 0.98
|
3.43 ± 1.01
|
2.30 ± 0.45
|
1.80 ± 0.37
|
1.28 ± 0.34
|
<0.001
|
HbA1c (%)
|
5.77 ± 0.41
|
5.75 ± 0.41
|
5.75 ± 0.41
|
5.76 ± 0.41
|
5.82 ± 0.42
|
<0.001
|
Sex, n (%)
|
|
|
|
|
|
<0.001
|
Male
|
8793 (52.75)
|
2162 (57.01)
|
2128 (53.19)
|
2189 (52.36)
|
2314 (49.28)
|
|
Female
|
7877 (47.25)
|
1630 (42.99)
|
1873 (46.81)
|
1992 (47.64)
|
2382 (50.72)
|
|
Race, n (%)
|
|
|
|
|
|
<0.001
|
Non-Hispanic White
|
6998 (41.98)
|
2118 (55.85)
|
1837 (45.91)
|
1646 (39.37)
|
1397 (29.75)
|
|
Non-Hispanic Black
|
3619 (21.71)
|
504 (13.29)
|
647 (16.17)
|
842 (20.14)
|
1626 (34.63)
|
|
Mexican American
|
2897 (17.38)
|
503 (13.26)
|
725 (18.12)
|
849 (20.31)
|
820 (17.46)
|
|
Other
|
3156 (18.93)
|
667 (17.59)
|
792 (19.80)
|
844 (20.19)
|
853 (18.16)
|
|
Education level, n (%)
|
|
|
|
|
|
0.270
|
Less than high school
|
4947 (29.72)
|
1151 (30.43)
|
1236 (30.95)
|
1198 (28.70)
|
1362 (29.02)
|
|
High school or equivalent
|
3972 (23.86)
|
910 (24.05)
|
935 (23.41)
|
1000 (23.96)
|
1127 (24.01)
|
|
College or above
|
7725 (46.41)
|
1722 (45.52)
|
1823 (45.64)
|
1976 (47.34)
|
2204 (46.96)
|
|
PIR, n (%)
|
|
|
|
|
|
0.061
|
<1.3
|
4686 (30.95)
|
1049 (30.63)
|
1107 (30.36)
|
1180 (31.00)
|
1350 (31.66)
|
|
1.3-3.5
|
5900 (38.97)
|
1383 (40.38)
|
1439 (39.47)
|
1416 (37.20)
|
1662 (38.98)
|
|
≥3.5
|
4555 (30.08)
|
993 (28.99)
|
1100 (30.17)
|
1210 (31.79)
|
1252 (29.36)
|
|
Smoking status, n (%)
|
|
|
|
|
|
<0.001
|
Never
|
8595 (51.61)
|
1672 (44.12)
|
1982 (49.57)
|
2250 (53.88)
|
2691 (57.38)
|
|
Former
|
4712 (28.29)
|
1220 (32.19)
|
1161 (29.04)
|
1145 (27.42)
|
1186 (25.29)
|
|
Current
|
3347 (20.10)
|
898 (23.69)
|
855 (21.39)
|
781 (18.70)
|
813 (17.33)
|
|
Alcohol consumption, n (%)
|
10661 (68.42)
|
2425 (68.33)
|
2655 (71.10)
|
2667 (68.04)
|
2914 (66.56)
|
<0.001
|
Hypertension, n (%)
|
10799 (64.78)
|
2528 (66.67)
|
2539 (63.46)
|
2669 (63.84)
|
3063 (65.23)
|
0.012
|
CVD, n (%)
|
2277 (13.66)
|
794 (20.94)
|
531 (13.27)
|
497 (11.89)
|
455 (9.69)
|
<0.001
|
All-cause mortality, n (%)
|
2884 (17.33)
|
1126 (29.76)
|
708 (17.73)
|
549 (13.15)
|
501 (10.68)
|
<0.001
|
CVD mortality, n (%)
|
758 (4.55)
|
323 (8.52)
|
173 (4.32)
|
143 (3.42)
|
119 (2.53)
|
<0.001
|
Abbreviations: ALI, Advanced lung cancer inflammation index; BMI, Body Mass Index; ALT, Alanine Aminotransferase; AST, Aspartate Aminotransferase; BUN, Blood Urea Nitrogen; eGFR, estimated Glomerular Filtration Rate; NLR, Neutrophil-to-Lymphocyte Ratio; HbA1c, Glycated Hemoglobin; PIR, Poverty Income Ratio; CVD, Cardiovascular Disease.
Data are presented as mean ± standard deviation for continuous variables and n (%) for categorical variables. P-values were calculated using one-way ANOVA for continuous variables and chi-square test for categorical variables.
Kaplan–Meier analysis
The Kaplan-Meier analysis was used to initially differentiate the connection between all-cause mortality and CVD mortality in ALI and prediabetes individuals. A total of 2,884 fatalities from all causes were reported among the 16, 670 prediabetes patients, with 758 deaths being linked to CVD. The Kaplan-Meier survival curve analysis revealed substantial disparities in survival rates between the ALI quartile groups. As seen in Figure 2, patients with prediabetes who had greater ALI levels had lower all-cause and CVD mortality (P <0.0001, P <0.0001, respectively).
ALI and mortality
Table 2 shows the logistic regression model results, which show the relationship between the ALI and all-cause and CVD mortality in prediabetes patients. Regardless of whether variables were corrected, there was a significant negative relationship between ALI and all-cause and CVD mortality in patients with prediabetes. In the adjusted III model, compared to the lowest ALI quartile (Q1), higher ALI levels were associated with decreased all-cause mortality in prediabetes patients, The HR (95% CI) for the Quantile2, Quantile3, and Quantile4 groups were 0.78 (0.70-0.86), 0.73 (0.65-0.81), 0.70 (0.62-0.79), respectively (Ptrend < 0.001). A similar trend was also seen in CVD mortality, wherein patients with prediabetes who have greater levels of ALI also have lower CVD mortality. The HR (95% CI) for the Quantile2, Quantile3, and Quantile4 groups were 0.66 (0.54-0.81), 0.68 (0.55-0.85), 0.56 (0.44-0.73), respectively (Ptrend < 0.001).
Assessing the dynamic alterations in ALI was vital to the prognosis of individuals with prediabetes patients. The multivariable-adjusted HR for all-cause mortality was 0.95 (0.94-0.97) and for CVD mortality was 0.93 (0.90-0.96) for every 10U increment in ALI.
Table 2. Association between ALI and mortality outcomes in NHANES participants.
Exposure
|
Non-adjusted
|
Adjust I
|
Adjust II
|
Adjust III
|
All-cause mortality
|
|
|
|
|
ALI quartile
|
|
|
|
|
Q1
|
Reference
|
Reference
|
Reference
|
Reference
|
Q2
|
0.54 (0.49, 0.59)
|
0.74 (0.67, 0.81)
|
0.76 (0.68, 0.84)
|
0.78 (0.70, 0.86)
|
Q3
|
0.40 (0.36, 0.44)
|
0.70 (0.63, 0.78)
|
0.72 (0.64, 0.80)
|
0.73 (0.65, 0.81)
|
Q4
|
0.32 (0.29, 0.36)
|
0.64 (0.57, 0.72)
|
0.68 (0.60, 0.77)
|
0.70 (0.62, 0.79)
|
ALI (per 10 U increment)
|
0.86 (0.84, 0.87)
|
0.94 (0.93, 0.96)
|
0.95 (0.93, 0.96)
|
0.95 (0.94, 0.97)
|
P for trend
|
<0.001
|
<0.001
|
<0.001
|
<0.001
|
Cardiovascular mortality
|
|
|
|
|
ALI quartile
|
|
|
|
|
Q1
|
Reference
|
Reference
|
Reference
|
Reference
|
Q2
|
0.46 (0.38, 0.55)
|
0.66 (0.54, 0.80)
|
0.64 (0.53, 0.79)
|
0.66 (0.54, 0.81)
|
Q3
|
0.36 (0.30, 0.44)
|
0.72 (0.58, 0.88)
|
0.67 (0.54, 0.83)
|
0.68 (0.55, 0.85)
|
Q4
|
0.27 (0.22, 0.33)
|
0.58 (0.46, 0.73)
|
0.53 (0.42, 0.69)
|
0.56 (0.44, 0.73)
|
ALI (per 10 U increment)
|
0.84 (0.81, 0.86)
|
0.94 (0.91, 0.97)
|
0.93 (0.90, 0.96)
|
0.93 (0.90, 0.96)
|
P for trend
|
<0.001
|
<0.001
|
<0.001
|
<0.001
|
Data are presented as HR (95% CI). P<0.001
Abbreviations: ALI, Advanced lung cancer inflammation index; HR, hazard ratio; CI, confidence interval.
Non-adjusted model: unadjusted.
Adjust I model: adjusted for age, sex, race, education level, and poverty income ratio.
Adjust II model: adjusted for variables in Adjust I plus body mass index, smoking status, alcohol consumption, hypertension, and cardiovascular disease.
Adjust III model: adjusted for variables in Adjust II plus alanine aminotransferase, aspartate aminotransferase, blood urea nitrogen, creatinine, estimated glomerular filtration rate, and glycated hemoglobin.
All models used Cox proportional hazards regression with follow-up time as the time variable.
Non-linear relationships
The non-linear connection between ALI levels and prediabetes patients' long-term prognosis was depicted in Figure 3. An L-shaped non-linear connection was seen between prediabetes patients' all-cause and CVD mortality (all p < 0.001). The study identified that the inflection points for all-cause mortality and CVD mortality were 40.50 and 37.35, respectively. For ALI values below 40.50, an increase of 10 U ALI corresponded to a 26% reduction in the all-cause mortality risk (HR: 0.74, 95% CI: 0.69-0.80, Ptrend <0.0001). Yet, when ALI>40.50, the effect weakened (HR: 0.98, 95% CI: 0.96-1.00, Ptrend =0.0149). CVD mortality showed a similar trend, for ALI values below 37.35, an increase of 10 U ALI corresponded to a 32% reduction in the CVD mortality risk (HR: 0.68, 95% CI:0.57-0.80, Ptrend <0.0001). However, when ALI exceeded 37.35, every 10 U ALI increase led to a 4% decrease in CVD mortality risk (HR: 0.96, 95% CI:0.93-1.00, Ptrend =0.0248). The details were presented in Table 3.
Table 3. Threshold effect analysis of ALI on all-cause, CVD mortality in NHANES participants.
Model
|
All-cause mortality
|
P-Value
|
CVD mortality
|
P-Value
|
Model I
|
|
|
|
|
Linear effect
|
0.95 (0.94, 0.97)
|
<0.0001
|
0.93 (0.90, 0.96)
|
<0.0001
|
Model II
|
|
|
|
|
Knot (K)
|
4.050
|
|
3.735
|
|
Effect for ALI < K
|
0.74 (0.69, 0.80)
|
<0.0001
|
0.68 (0.57, 0.80)
|
<0.0001
|
Effect for ALI > K
|
0.98 (0.96, 1.00)
|
0.0149
|
0.96 (0.93, 1.00)
|
0.0248
|
Difference in effects
|
1.31 (1.21, 1.42)
|
<0.0001
|
1.42 (1.19, 1.70)
|
<0.0001
|
LR test p-value
|
<0.001
|
|
<0.001
|
|
Data are presented as HR (95% CI) for each 10 U increase in ALI.
CVD: Cardiovascular disease; LR: Likelihood ratio
Models were adjusted for age, sex, race/ethnicity, education level, poverty income ratio, body mass index, smoking status, alcohol consumption, hypertension, alanine aminotransferase, aspartate aminotransferase, blood urea nitrogen, creatinine, estimated glomerular filtration rate, glycated hemoglobin, and history of cardiovascular disease.
Sensitivity analysis
Sensitivity analysis was carried out using stratified analysis and interaction analysis. In order to evaluate possible interactions between stratified factors and ALI, stratified analyses for all-cause and CVD mortality among patients with prediabetes were shown in Tables S1 and S2. There were no significant interactions seen in either all-cause or CVD mortality. Furthermore, Tables S3 and S4 showed a stratified study of all-cause and CVD mortality in prediabetes patients, grouped by inflection points, to investigate potential interactions between stratified factors and ALI. There were no significant interactions observed. Given that interaction analyses revealed no statistically significant differences across all subgroups (all P for interaction >0.05), indicating the consistency and robustness of the inverse relationship between ALI and mortality risk across diverse demographic and clinical subgroups.