Skip to main content

Table 3 Threshold effect analysis of the relationship between all-cause or CVD mortality and HALP score in participants with diabetes or prediabetes

From: Association between Hemoglobin–albumin–lymphocyte–platelet score and all-cause or cardiovascular mortality in patients with diabetes or prediabetes: mediated effects of renal function

  

Adjusted HR(95%CI)

P

All-cause mortality

   

 HALP

   

 Model 1

One-step cox regression

0.997(0.995–0.999)

< 0.004

 Model 2

Two-piecewise cox regression

  

HALP < 52

0.964(0.957–0.971)

< 0.001

HALP ≥ 52

1.001(1.000-1.002)

0.153

Likelihood ratio test

 

< 0.001

CVD mortality

   

 HALP

   

 Model 1

One-step cox regression

0.992(0.987–0.996)

< 0.001

 Model 2

Two-piecewise cox regression

  

HALP < 49

0.940 (0.927–0.953)

< 0.001

HALP ≥ 49

1.000(0.997–1.003)

0.955

Likelihood ratio test

 

< 0.001

  1. Model 1 and Model 2: adjusted for gender, race, age, drinking status, smoking status, education levels, married status, with or without HTN, BMI, PIR, TG, Chol, and HDL-C
  2. HALP = Hemoglobin-albumin-lymphocyte-platelet, HTN = hypertension, BMI = body mass index, PIR = poverty income ratio, HbA1c, glycosylated hemoglobin; TG, triglyceride; Chol, cholesterol, HDL-C, high-density lipoprotein- cholesterol