Clinical Nutrition xxx (2014) 1e6

Contents lists available at ScienceDirect

Clinical Nutrition journal homepage: http://www.elsevier.com/locate/clnu

Original article

Association between magnesium status, oxidative stress and inflammation in preeclampsia: A caseecontrol study Vivianne de Sousa Rocha a, 1, Fernanda Brunacci Della Rosa a, Rodrigo Ruano b, c,
lia Colli a, * Marcelo Zugaib b, Ce ~o Experimental, Faculdade de Ci^ ~o Paulo, Av. Professor Lineu Prestes 580, Departamento de Alimentos e Nutriça encias Farmac^ euticas, Universidade de Sa ~o Paulo, SP, Brazil Bloco 14, 05508-000 Sa b ~o Paulo, Av. Dr. Arnaldo 455, Cerqueira C
~o Paulo, SP, Brazil Departamento de Obstetrícia, Faculdade de Medicina, Universidade de Sa esar, 01246903 Sa c Department of Obstetrics and Gynecology, Baylor College of Medicine, Houston, TX, USA a

a r t i c l e i n f o

s u m m a r y

Article history: Received 15 May 2014 Accepted 1 December 2014

Background & aims: Preeclampsia is responsible for more than one-third of all maternal deaths in Brazil. The objectives of the present study were to evaluate magnesium status and its association with oxidative stress and inflammation in preeclamptic women, and to identify the predictor variables of the disorder. Methods: The study population consisted of 36 women divided into preeclamptic (n ¼ 18) and control groups (n ¼ 18). The preeclamptic group included women (20 weeks of pregnancy) with arterial pressure  140/90 mmHg and proteinuria >0.3 g/24 h, while the control group comprised pregnant women with no clinical/obstetric complications. Magnesium intake was assessed via a food frequency questionnaire validated for pregnant women in Brazil. Plasma, erythrocyte and urinary magnesium levels were determined by flame atomic absorption spectroscopy, while oxidative stress and inflammatory markers were assessed using standard protocols. Logistic regression analysis was used to identify the predictors of preeclampsia. Results: Preeclamptic and control groups were similar with respect to magnesium intake and urinary excretion, while plasma and erythrocyte magnesium concentrations were higher in the former group. Plasma magnesium was positively correlated with catalase and glutathione peroxidase activities and with concentrations of interleukin-6 and tumor necrosis factor alpha. Regression analysis showed that plasma magnesium and urinary 8-isoprostane were associated with preeclampsia. Conclusion: Magnesium status appears to result from homeostatic imbalance and physiological alterations typical of preeclampsia. Increased plasma magnesium and decreased urinary 8-isoprostane were considered predictors of preeclampsia. © 2014 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

Keywords: Magnesium Preeclampsia Oxidative stress Inflammation

1. Introduction

Abbreviations: AUC, area under the curve; CAT, catalase; CI, confidence interval; CRP, C-reactive protein; D, difference between magnesium intake and EAR; EAR, estimated average requirement; FFQ, food frequency questionnaire; GSH-Px, glutathione peroxidase; Hb, hemoglobin; HPLC, high performance liquid chromatography; IL6, interleukin 6; MDA, malondialdehyde; NF, nuclear factor; OR, odds ratio; PCTL, percentile; RDA, recommended dietary allowance; ROC, receiver operating characteristic; ROS, reactive oxygen species; SDD, standard deviation of D; TBA, thiobarbituric acid; TNF-a, tumor necrosis factor alpha. * Corresponding author. Tel.: þ55 11 3091 3651. E-mail address: [email protected] (C. Colli). 1 Present address. Núcleo de Nutriç~ ao, Universidade Federal de Sergipe, Campus

rio Professor Anto ^nio Garcia Filho, Rua Padre Alvares Universita Pitangueira 248, Centro, 49400-000 Lagarto, SE, Brazil.

Preeclampsia is one of the main causes of maternal and fetal mortality worldwide [1], while in Brazil, preeclampsia and eclampsia are responsible for 37% of all maternal deaths [2]. For reasons that are not fully understood, the placenta does not function properly in preeclamptic women resulting in oxidative stress, exacerbated maternal inflammation and emergence of classical symptoms such as increased arterial blood pressure and proteinuria [3]. Although the physiopathological mechanisms associated with the disorder have been studied extensively, the influence of factors such as magnesium status remains obscure. Some researchers have associated preeclampsia with the deficiency of magnesium since

http://dx.doi.org/10.1016/j.clnu.2014.12.001 0261-5614/© 2014 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

Please cite this article in press as: de Sousa Rocha V, et al., Association between magnesium status, oxidative stress and inflammation in preeclampsia: A caseecontrol study, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.12.001

2

V. de Sousa Rocha et al. / Clinical Nutrition xxx (2014) 1e6

this mineral is required for numerous biochemical functions including energy production, the modulation of vascular tone and muscle contraction [4e6]. Magnesium deficiency provoked by reduced dietary intake is characterized by alterations in the compartmental distribution of the micronutrient [6], which may result in the generation of reactive oxygen species (ROS) and associated inflammation [7,8]. However, there are currently no reports concerning the relationships between magnesium status and oxidative stress, inflammation and endothelial functions in preeclampsia, although these factors appear to be critical for the development of the condition. The objective of the present study was to test the hypothesis that magnesium deficiency aggravates oxidative stress and inflammatory response in preeclampsia. For this purpose, magnesium intake and the concentrations of the micronutrient in plasma, erythrocytes and urine were evaluated in groups of preeclamptic and healthy pregnant women, and associations between magnesium status and markers of oxidative stress and inflammatory cytokines were determined in order to identify predictor variables of preeclampsia. 2. Materials and methods 2.1. Study population Details of the project were submitted to and approved by the Ethical Research Committees of the Hospital das Clínicas, Faculdade ^ncias Farmace ^uticas, Unide Medicina and Faculdade de Cie ~o Paulo (protocol CAAE #1105.0.015.018-09) and versidade de Sa ~o Paulo, SP, authorized by the Director of Hospital Ipiranga (Sa Brazil). Written informed consent was obtained from all participants prior to the commencement of the study. The sample size was calculated from the differences between means (in terms of standard deviations) of the two groups using Student t test, considering an effect size (Cohen's d) equivalent to 1 [9], assuming a statistical difference between the groups of 5% and a study power of 80%. The minimum sample size required was estimated to be 34, i.e. 17 subjects per group. A total of 36 pregnant women (18 preeclamptic subjects and 18 healthy controls), all of whom had attended the obstetric services of Hospital Ipiranga and Hospital das Clínicas da Faculdade de ~o Paulo, Sa ~o Paulo, SP, Brazil), were Medicina (Universidade de Sa recruited consecutively over a period of 33 months according to predefined inclusion and exclusion criteria. For every preeclamptic individual recruited, a further healthy pregnant woman of similar gestational age was incorporated into the control group. Preeclampsia was diagnosed on the basis of maternal blood pressure 140/90 mmHg and proteinuria >0.3 g/24 h [1] after the 20th week of pregnancy. The control group included pregnant women of similar gestational ages but with no diagnosis of preeclampsia. The exclusion criteria for both groups were women presenting chronic arterial hypertension, gestational hypertension (detected before the 20th week of pregnancy), multiple gestation, type I or type II diabetes, heart disease, smoking addiction or use of a mineral supplement containing magnesium. All participants were advised to ingest iron sulfate supplement as recommended by the Brazilian Ministry of Health. 2.2. Study design This observational caseecontrol study involved a group of pregnant women selected initially on the basis of information contained in medical records. At the first meeting with potential participants, the aims and objectives of the investigation were carefully explained and subjects were invited to take part in the

study and to sign the informed consent. Demographic data (age, number of pregnancies and family history of preeclampsia) were collected using a questionnaire, and the information provided was verified by comparison with medical records. Participants were instructed regarding the preconditions for sampling blood after 8 h fasting and the procedure for collecting a 24-h urine sample. At the second meeting, participants were submitted to anthropometric assessment and a quantitative food frequency questionnaire (FFQ) was applied. During this meeting, urine samples collected by participants during the previous 24-h period were received and stored at 80  C until required for analysis, and blood samples (20 mL) were obtained by venipuncture and stored in tubes with and without anticoagulant (10 mL sodium citrate/mL blood or 7.2 mg of ethylenediaminetetraacetic acid). 2.3. Anthropometric evaluation Anthropometric measurements were performed with the subject barefoot and wearing light clothing. Body weight was deter~o Paulo, Brazil) of mined using portable digital scales (Toledo, Sa 250 kg capacity and 50 g sensitivity, while height was evaluated using a Toledo stadiometer, comprising a 2 m measuring tape graduated in 0.5 cm divisions, with the individual in an orthostatic position. Maternal nutritional status was based on the body mass index (BMI) for the week of pregnancy [10], as recommended by the Brazilian Ministry of Health. 2.4. Assessment of magnesium intake An FFQ, validated for use with pregnant women in Brazil [11], was applied by a trained nutritionist and the data were analyzed using the on-line NutriQuanti system [12]. Magnesium intake was assessed according to dietary reference intakes [13], while the probability of nutrient inadequacy was determined by calculating the difference (D) between the observed intake of the individual and the estimated average requirement according to age and physiological status. The standard deviation of D (SDD) was calculated and the D/SDD ratio used to determine the probability of correctly concluding that the magnesium intake value of an individual was adequate or inadequate in relation to the mean reference value [14]. 2.5. Biochemical analyses Quantitative determinations of the levels of magnesium in serum, plasma and urine were performed using a flame atomic absorption spectrometer (AAnalyst 100, PerkineElmer, Norwalk, CT, USA). All glassware and tools were subjected to demineralization with 30% nitric acid for 12 h prior to use in assays. In order to reduce potential interference by aluminum, a solution of 5% lanthanum oxide was added to all samples, including those employed in constructing the calibration curve, to give a final concentration of lanthanum of 0.1% [15]. Standard reference samples (Seronorm™ Trace Elements Serum L-1 and Urine Blank, Sero, Billingstad, Norway), together with secondary reference samples (plasma, erythrocyte and urine pools), were used to determine the performance of the analytical procedures, revealing an accuracy of 101% and a precision between 84 and 99%. Plasma malondialdehyde (MDA) concentration, taken as a measure of lipid peroxidation, was assessed by high performance liquid chromatography (HPLC) performed using a Shimadzu (Kyoto, Japan) instrument fitted with a reverse-phase C18 column (150  4.6 mm; 5 mm particle size; Phenomenex, Torrance, CA). Reaction between MDA and thiobarbituric acid (TBA) formed the MDA-TBA2 chromogen, the absorbance of which was monitored at

Please cite this article in press as: de Sousa Rocha V, et al., Association between magnesium status, oxidative stress and inflammation in preeclampsia: A caseecontrol study, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.12.001

V. de Sousa Rocha et al. / Clinical Nutrition xxx (2014) 1e6

532 nm. A calibration curve was constructed for MDA concentrations in the range 0.1e10 mM [15]. Urinary 8-isoprostane concentrations were determined using an enzyme-linked immunosorbent assay kit (8-Isoprostane EIA Kit Cat. no. 516351; Cayman Chemical Co, Ann Arbor, MI, USA). Urine samples were diluted in deionized water in the proportion 1:20 and assays were carried out according to the recommendations of the manufacturer. Absorbances were recorded at 412 nm using a SpectraMax® microplate reader (Molecular Devices Corporation, Sunnyvale, CA, USA). The activities of glutathione peroxidase (GSH-Px) and catalase (CAT) were evaluated in samples of lysed erythrocytes. GSH-Px activity was assessed indirectly via a coupled reaction in which oxidized glutathione (produced by the action of GSH-Px on an organic hydroperoxide) was reduced by glutathione reductase and NADPH. The reaction mixture was incubated for 6 min at 37  C and the oxidation of NADPH to NADP was monitored at 340 nm [16,17]. CAT activity was measured by the decomposition of hydrogen peroxide in TriseHCl buffer (pH 8.0) at 37  C and monitored for 6 min at 230 nm [18]. Enzyme activities were expressed as U/g hemoglobin (U/g Hb). The concentration of Hb in packed erythrocytes was assessed using the cyanmethemoglobin method with the
stica®, Lagoa Santa, MG, aid of a commercial kit (Labtest Diagno Brazil). Serum C-reactive protein (CRP), interleukin 6 (IL-6) and tumor necrosis factor alpha (TNF-a) were analyzed using, respectively, Human C-reactive Protein Immunoassay (Cat. no. DCRP00), Human IL-6 Immunoassay (Cat. no. S6050) and Human TNF-a Immunoassay (Cat. no. STA00C) kits from R&D Systems, Millipore Corporation, Billerica, MN, USA. Absorbances were determined using a SpectraMax® microplate reader (Molecular Devices Corporation). 2.6. Statistical analysis All analyses were performed using R software version 2.14.2 with the significance level (a) set at 5%. Dietetic data (nutrient intake) were adjusted for total energy intake using the residual method in order to reduce errors inherent in food measurements [19]. Assessment of normality of distribution of data was performed using ShapiroeWilk test. In order to compare participants in the two groups, c2 tests with Yates correction and ManneWhitney tests were employed. The associations between the variables were determined using the non-parametric Kendal t correlation. Logistic regression analysis was performed to identify predictor variables that distinguished most clearly between the preeclamptic and control groups. The results were expressed as odds ratios (OR) and 90% confidence intervals (CI90%). The performance of the adjusted logistic regression model was evaluated by means of the receiver operating characteristic (ROC) curve, considering the area under the curve (AUC) as representing the probability of subjects being correctly ranked, through determination of sensitivity (true positives) and specificity (true negatives). The quality of the adjusted model was verified using envelope plots and Cook's distance. 3. Results There were no differences between the preeclamptic and control groups with respect to age and gestational week, as shown by the median values and the 25th and 75th percentiles (Table 1). The groups were also similar regarding magnesium intake (Table 2), but none of the women showed a probability of adequacy of magnesium intake greater than the 85th percentile (Fig. 1). Although the preeclamptic and control groups presented similar levels of urinary magnesium, the concentrations of plasma and

3

Table 1 Demographic and clinical characteristics of the study groups. Variablea

Control group (n ¼ 18)

Age [years; median (PCTLs)] Gestational weeks [median (PCTLs)] Systolic arterial pressure [mmHg; median (PCTLs)] Diastolic arterial pressure [mmHg; median (PCTLs)] Proteinuria [g/24 h; median (PCTLs)] Family history of preeclampsia No Yes Not known Parturitions [n (%)] Nulliparous Primiparous or multiparous Gestational weight [kg; median (PCTLs)] Height [m; median (PCTLs)] Body mass index [kg/m2); median (PCTLs)] Nutritional status [n (%)] Underweight Eutrophic Overweight Obese

27 (24; 32) 22 (18; 30) 31.5 (28.8; 33.3) 31.5 (26.8; 35.0)

Preeclamptic group P (n ¼ 18) 0.091b 0.998b

103 (100; 110)

160 (150; 161)