Assessment of preeclampsia risk by use of serum ionized magnesium-based equation

Figures & data

Table 1. Characteristics of women with normal pregnancy and pregnancy with preeclampsia.

Clinical variables	Normal pregnancy	Pregnancy with preeclampsia	p value
Number (n)	64	20
Maternal age (years ±SD)	28.14 ± 6.8	29.50 ± 9.0	.48
Normal maternal-age pregnancy (age 20–34 years, n = 56)	80.4%	19.6%
Teenage pregnancy (age ≤19 years, n = 9)	66.6%	33.3%	.39*
Elderly pregnancy (age ≥35 years, n = 19)	68.4%	31.6%	.35*
Gestational age at delivery (week)	39.0 ± 1.2	38.0 ± 1.3	.07
Low birth weight (baby birth weight <2500 g)	1.9%	35.0%	<.001
Number of delivery by Cesarean section (%)	33.3%	80.0%	<.001
Primigravida (n = 42, %)	69.0%	31%	.12**
Baseline serum creatinine (mg/dL ±SD)	0.52 ± 0.1	0.56 ± 0.1	.09

* p values compare to normal maternal-age pregnancy.

** p values compare to multiparity.

Figure 1. The box plots of the mean and SD of laboratory results: they demonstrated total magnesium levels (Picture A), ionized magnesium levels (Picture B), ionized magnesium fraction (%) (Picture C) and serum albumin levels (Picture D) during normal pregnancy and pregnancy with preeclampsia. Only ionized magnesium fraction and serum albumin levels showed the significant different between normal pregnancy and preeclampsia.

Figure 2. Curve of estimation on disease probability: the graph showed association between the probability of preeclampsia and the ionized magnesium fraction. The maximum slope indicated a cut point value of 24.67%, meaning that patients with ionized magnesium fractions of less than 24.67% were at higher risk of preeclampsia.

Table 2. Association between preeclampsia and magnesium, teenage and elderly primigravida.

	Unadjusted OR (95% CI)	Adjusted OR* (95% CI)	p value (for adjust OR)
The ionized magnesium fraction	4.41 (1.46, 13.40)	5.55 (1.66, 18.64)	.006
(<24.67%)
Teenage primigravida	5.47 (1.85, 35.42)	8.64 (1.179, 63.308)	.034
Elderly primigravida	11.11 (1.09, 113.78)	16.90 (1.48, 193.39)	.023

* The results of multivariate regression analysis also demonstrated the Ombinus p values = .001 with −2.021 of constant equation value. The coefficient values were 1.715, 2.157 and 2.827 for ionized magnesium fraction, teenage primigravida and elderly primigravida, respectively.

Therefore, the logistic equation = 1.715 (the ionized magnesium fraction) + 2.157(teenage primigravida) + 2.827(elderly primigravida) − 2.021.

Figure 3. The ROC of the predictive accuracy of an ionized magnesium fraction-based equation model for preeclampsia: the ionized magnesium fraction-based equation model was derived from the logistic regression analysis on the fraction, teenage as well as elderly primigravida. The area under ROC curve was 0.77, indicating a significant degree of discrimination (p < .001). Scores of >0.27 were highly suggestive of preeclampsia with 70% sensitivity and 81% specificity.