Table 3 Regression models predicting... | American Society of Hematology

Table 3

Regression models predicting fractional (% of dose) and absolute (mg) iron absorption from iron supplements in relation to timing of administration, dosage, body iron stores, and hepcidin concentrations at administration or at start of the study

	Dependent variable
	Log absorption (%), n = 98						Log absolute absorption mg, n = 98
	Independent variables, standardized β						Independent variables, standardized β
Model R²*	Time	Dose (mg)	BIS mg/kg BW	Log PHep nM	BIS start mg/kg BW	Log PHep, start nM	Model R²	Time	Dose (mg)	BIS mg/kg BW	Log PHep nM	BIS start mg/kg BW	Log PHep, start nM
0.689	−0.146†	−0.352‡	−0.463‡	−0.206§	—	—	0.659‖	−0.179§	0.656‡	−0.337‡	−0.222§	—	—
0.666	−0.165†	−0.371‡	−0.593‡	—	—	—	0.633¶	−0.199§	0.636‡	−0.477‡	—	—	—
0.579	−0.228§	−0.397†	—	−0.486†	—	—	0.601#	−0.238§	0.623‡	—	−0.426‡	—	—
0.604	−0.367‡	−0.319‡	—	—	−0.538‡	—	0.629¶	−0.346‡	0.679‡	—	—	−0.488‡	—
0.520	−0.367‡	−0.529‡	—	—	—	−0.423‡	0.599#	−0.341‡	0.516‡	—	—	—	−0.363‡
0.378	−0.385‡	−0.514‡	—	—	—	—	0.446**	−.376‡	0.521‡	—	—	—	—

	Dependent variable
	Log absorption (%), n = 98						Log absolute absorption mg, n = 98
	Independent variables, standardized β						Independent variables, standardized β
Model R²*	Time	Dose (mg)	BIS mg/kg BW	Log PHep nM	BIS start mg/kg BW	Log PHep, start nM	Model R²	Time	Dose (mg)	BIS mg/kg BW	Log PHep nM	BIS start mg/kg BW	Log PHep, start nM
0.689	−0.146†	−0.352‡	−0.463‡	−0.206§	—	—	0.659‖	−0.179§	0.656‡	−0.337‡	−0.222§	—	—
0.666	−0.165†	−0.371‡	−0.593‡	—	—	—	0.633¶	−0.199§	0.636‡	−0.477‡	—	—	—
0.579	−0.228§	−0.397†	—	−0.486†	—	—	0.601#	−0.238§	0.623‡	—	−0.426‡	—	—
0.604	−0.367‡	−0.319‡	—	—	−0.538‡	—	0.629¶	−0.346‡	0.679‡	—	—	−0.488‡	—
0.520	−0.367‡	−0.529‡	—	—	—	−0.423‡	0.599#	−0.341‡	0.516‡	—	—	—	−0.363‡
0.378	−0.385‡	−0.514‡	—	—	—	—	0.446**	−.376‡	0.521‡	—	—	—	—

BIS, body iron stores; BW, body weight; PHep, plasma hepcidin.

Differences between nested models were tested with change in F statistic; differences between non-nested models were tested by comparing different R coefficients with the Steiger Z test. Reported models differ significantly if superscript differs, P < .01. All models are significant at P < .01. Differences between nested models were tested with change in F statistic; Differences between non-nested models were tested by comparing different R coefficients with the Steiger Z test. Within one category of dependent variables, reported models differ significantly if superscript differs, P < .01.

Different from other models of log absorption P < .01.

†

Significant regression parameter P < .05.

‡

Significant regression parameter P < .001.

Significant regression parameter P < .01.

‖

Different from other models P < .01.

Different from other models, P < .01.

Different from all models, P < .01.

View Large

This Feature Is Available To Subscribers Only