Hostname: page-component-586b7cd67f-2plfb Total loading time: 0 Render date: 2024-11-25T07:08:15.990Z Has data issue: false hasContentIssue false

Depressive symptoms among Mexican adolescent girls in relation to iron status, anaemia, body weight and pubertal status: results from a latent class analysis

Published online by Cambridge University Press:  18 May 2022

Arli Guadalupe Zarate-Ortiz*
Affiliation:
Division of Human Nutrition and Health, Wageningen University & Research, Wageningen 6700 AK, The Netherlands
Hans Verhoef
Affiliation:
Division of Human Nutrition and Health, Wageningen University & Research, Wageningen 6700 AK, The Netherlands Medical Research Council (MCR) Unit, The Gambia at London School of Hygiene & Tropical Medicine, Serrekunda, Gambia
Alida Melse-Boonstra
Affiliation:
Division of Human Nutrition and Health, Wageningen University & Research, Wageningen 6700 AK, The Netherlands
Bo-Jane Woods
Affiliation:
Division of Human Nutrition and Health, Wageningen University & Research, Wageningen 6700 AK, The Netherlands Access to Nutrition Initiative, Utrecht, The Netherlands
Elida Estefania Lee-Bazaldúa
Affiliation:
Department of Nutrition, Universidad de Monterrey, San Pedro Garza García, Mexico
Edith JM Feskens
Affiliation:
Division of Human Nutrition and Health, Wageningen University & Research, Wageningen 6700 AK, The Netherlands
Angelica Quiroga-Garza
Affiliation:
Department of Psychology, Universidad de Monterrey, San Pedro Garza García, Mexico
Ana Carla Cepeda-Lopez
Affiliation:
Department of Nutrition, Universidad de Monterrey, San Pedro Garza García, Mexico
*
*Corresponding author: Email [email protected]
Rights & Permissions [Opens in a new window]

Abstract

Objective:

The study examined the association between depressive symptoms and iron status, anaemia, body weight and pubertal status among Mexican adolescent girls.

Design:

In this cross-sectional study, depressive symptoms were assessed by the 6-item Kutcher Adolescent Depression Scale, and latent class analysis (LCA) was used to identify and characterise groups of girls based on depressive symptoms. Iron status and inflammation were assessed using ferritin and soluble transferrin receptor, C-reactive protein and alpha-1-acid glycoprotein, respectively. Multiple logistic and linear regressions were applied to model class membership as a function of iron status, anaemia, body weight and pubertal status.

Participants:

We collected data from 408 girls aged 12–20 years.

Setting:

Public schools in northern Mexico.

Results:

LCA yielded three classes of depressive symptoms: 44·4 % of the adolescents were ‘unlikely to be depressed’, 41·5 % were ‘likely to be depressed’ and 14·1 % were ‘highly likely to be depressed’. Our analyses demonstrated that iron-deficient girls had greater odds of being ‘likely depressed’ (OR 2·01, 95 % CI 1·01, 3·00) or ‘highly likely depressed’ (OR 2·80, 95 % CI 1·76, 3·84). Linear regression analyses revealed that lower Hb concentrations and higher body weight increased the probability of being ‘likely depressed’. There was no evidence that depressive symptoms were associated with age at menarche and years since menstruation.

Conclusions:

This study shows that iron-deficient adolescent girls are more likely to suffer from depressive symptoms and that lower concentrations of Hb and higher body weight increased the probability of experiencing depressive symptoms.

Type
Research Paper
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
© The Author(s), 2022. Published by Cambridge University Press on behalf of The Nutrition Society

Previous observational studies have reported an association between iron deficiency and depressive symptoms in children and adolescents(Reference Campisi, Zasowski and Shah1). Adolescence is a period of intensive brain remodelling, and iron has a role in various neurological functions like myelin production, synaptogenesis and production of neurotransmitters, that is serotonin, norepinephrine and dopamine(Reference Georgieff2,Reference Gholamreza Noorazar, Ranjbar and Nemati3) . Disrupted myelination and lower concentration of these neurotransmitters are common in depressed persons(Reference Saluja, Iachan and Scheidt4,Reference Webb, Nishida and Darnton-Hill5) . Randomised controlled trials in women with postpartum depression have shown that iron supplementation improves depressive symptoms in iron-deficient and non-iron-deficient women(Reference Wassef, Nguyen and St-André6). In addition, low Hb concentration has been associated with an increased odds of adult depression by 43 % (95 % CI 1·23, 1·65)(Reference Saluja, Iachan and Scheidt4). Anaemia and depression share symptoms such as fatigue, which may be the result of altered cerebral oxygen transportation. However, little attention has been paid to the role of iron deficiency and anaemia in depression among adolescents, especially girls who recently started their menses.

The low-degree chronic inflammation produced by adiposity might also represent a risk factor for depression. Longitudinal studies showed that BMI at the age of 14 years predicted BMI and inflammatory markers at the age of 17 years, where inflammation was also associated with depressive symptoms(Reference Oddy, Allen and Trapp7). In addition, depressive symptoms were associated with increased IL-6 responses among adolescents with higher BMI but not among those with lower BMI(Reference Chiang, Bower and Irwin8). The link between high BMI and depressive symptoms might not be explained exclusively through the biologically plausible pathways of adiposity and inflammation but also by psychosocial pathways of self-esteem, body dissatisfaction and social support(9Reference Colich, Kircanski and Foland-Ross11). Furthermore, we showed earlier that overweight and obesity increase the risk for iron deficiency(Reference Cepeda-Lopez, Osendarp and Melse-Boonstra12).

After pubertal onset, prevalence of depression in girls is approximately twice that of boys(Reference Vanes, Moutoussis and Ziegler13Reference Lee and Kim15). In a survey among Mexican high school students, depressive symptoms were reported by 34 % and 18 % of female and male students, respectively(9). Pubertal development may be the mechanism that underlies the gender difference in rates of depression. Clinical and epidemiological evidence suggests that fluctuations in hormonal concentrations, particularly in oestrogens, may influence the regulation of the hypothalamic–pituitary–adrenal axis and this may alter the neurotransmitter systems(Reference Deecher, Andree and Sloan10,Reference Shariatpanaahi, Shariatpanaahi and Moshtaaghi16) . Anomalous hypothalamic–pituitary–adrenal axis function has been associated with the onset of depression in adolescents, and this effect seems to be dependent on the pubertal stage(Reference Colich, Kircanski and Foland-Ross11). In addition, the rise of androgens during puberty, most notably dihydrotestosterone and testosterone, is involved in the hippocampus development, and larger hippocampal volume is associated with the risk of depression(Reference Ellis, Fernandes and Simmons17). Thus, puberty is a critical period of development in which examining factors associated with depression is essential.

This study aimed to examine the association between iron status and depressive symptoms in Mexican adolescent girls. In addition, we also explored whether Hb concentration, body weight and pubertal onset were associated with depressive symptoms.

Methods

Participants

We conducted a cross-sectional study in the cities of Santa Catarina and Monterrey, in Northern Mexico, from September 2018 to January 2019. Adolescent girls aged 12–20 years from public schools were recruited, and written informed consent was obtained from the adolescents and their parents. Initially, we planned to conduct an intervention study among 162 iron-deficient participants. Since the prevalence of iron deficiency in Mexican adolescent girls was 36 % in urban areas, we determined that we would need to screen 450 adolescent girls to find a large enough number of iron-deficient participants for our intervention study. Hence, this dictated the sample size of this study. Exclusion criteria were diagnosis of systemic disease that may affect iron status, history of major surgery in the last month and regular use of medication (except contraceptives). After applying these criteria, a total of 408 adolescent girls were recruited for complete assessment. Five girls had incomplete questionnaires and therefore were excluded from this analysis.

Measurements

Depressive symptoms

In the remainder of this paper, ‘depressive symptoms’ refer to symptoms experienced by adolescents, such as depressed mood, loss of interest, reduced energy leading to increased fatigability and diminished activity. To assess depressive symptoms, we used the 6-item Kutcher Adolescent Depression Scale (6-KADS)(Reference LeBlanc, Almudevar and Brooks18) (Appendix 1), a questionnaire that has been translated from English to Spanish and has been validated for use in Latino adolescents(Reference Romero-Acosta, Gómez-de-Regil and Lowe19). The questionnaire was applied individually and as privately as possible by trained staff members. Six questions briefly describe the characteristic symptomatology of depression, with a Likert scale of four points (0 = ‘almost never’ to 3 = ‘all the time’). The overall score consisted of the sum of the score of all the items. Therefore, the total score ranges between 0 and 18. Traditionally, individuals are classified as not showing depressive symptoms if the overall score is 0–6 and as showing depressive symptoms if the score is ≥ 6.

Iron status

A trained staff member drew venous blood samples for subsequent assessment of iron indicators (serum concentrations of ferritin and soluble transferrin receptor) and inflammatory markers (serum concentrations of C-reactive protein and alpha-1-acid glycoprotein). Blood samples were drawn at school facilities between 08.00 and 12.00 hours approximately; not all the participants had fasted; however, fasting has no implications on ferritin concentrations. Serum samples were stored at –80°C. Iron and inflammation concentrations were measured using a sandwich ELISA by the VitMin Lab, Willstaett, Germany(Reference Erhardt, Estes and Pfeiffer20). Depleted iron stores were defined as serum concentrations of ferritin < 15 µg/l(21). The values for soluble transferrin receptor reported by the VitMin Lab are in the same range as the RAMCO assay. Therefore, iron-deficient erythropoiesis was defined as soluble transferrin receptor concentration > 8·3 mg/l(Reference Erhardt, Estes and Pfeiffer20).

Anaemia

Anaemia was defined by capillary blood Hb concentrations < 120 g/l(22), measured by a HemoCue 201+ portable photometer (HemoCue AB.).

Anthropometric measurements

Body weight was measured on a calibrated platform scale with a bar (SECA 700), to the nearest 100 g, and with participants wearing light clothes. Height was measured in centimetres with the subject barefoot using a stadiometer (SECA 213). Both body weight and height were measured in duplicate and using standardised techniques. A third measure was done if the two measures differ by more than 500 g or 1 cm, respectively. BMI-for-age (BAZ) was used to classify weight status according to the WHO growth standards(23). BAZ was categorised as obese (> 2 sd), overweight (> 1 sd), normal (–2 sd to 1 sd) and thin (< –2 sd).

Age at menarche

Self-reported age at menarche (first menstruation) was used as a proxy indicator for the onset of puberty.

Years since menarche

Self-reported age at menarche was subtracted from the chronological age.

2Data analysis

The statistical programme IBM SPSS 27.0 was used to calculate the descriptive statistics of general characteristics of the study population. Latent GOLD 5.1 (https://www.statisticalinnovations.com/latent-gold-5-1/) was used to undertake the latent class analysis (LCA). LCA is a statistical technique that aims to identify distinctive subgroups of people who share common characteristics so that people within the same subgroup have a similar scoring pattern on the measured variables(Reference Kongsted and Nielsen24). To model the association between the different classes of depressive symptoms and the exposure variables (iron status, anaemia, body weight and pubertal onset), three main steps were involved(Reference Vermunt and Magidson25): (1) estimating the model with the optimal number of latent classes; (2) classifying the adolescent girls into one of the classes based on the model selected in step 1; and (3) examining the relationship between the classes and the exposure variables.

Responses to 6-KADS items from 403 Mexican adolescent girls were available and complete for LCA. First, explanatory LCA with 1–5 classes was conducted with the six 6-KADS items introduced as ordinal indicators. We selected the model with the lowest value for the Bayesian information criterion, which indicates the balance between model fit and model simplicity. It is a basic assumption from LCA that external variables are not correlated within the identified classes, which is known as conditional independence(Reference Kongsted and Nielsen24). After identifying the optimal class solution, we examined the conditional independence between latent class indicators and the exposure variables by inspecting the bivariate residuals after including each exposure variable one by one (see online supplementary material, Supplemental Table 1). Bivariate residual values higher than 3·0 indicate a residual association between variables(Reference Vermunt and Magidson25). We concluded that BMI-for-age was correlated with item 3 and item 6 of the 6-KADS questionnaire. We re-estimated the model, including BMI-for-age as an active covariate, thus correcting the encountered effect between BMI-for-age and items 3 and 6. This correction resulted in a reduction in the Bayesian information criterion value. Second, we used posteriori classification to assign each of the adolescent girls to one of the three latent classes from the corrected model, and this classification information was saved. Third, we used two types of exposure variables, nominal (for iron deficiency and iron-deficient erythropoiesis) and continuous (for Hb, BMI z-scores, years since menstruation and age at menarche), which we modelled using a multinomial logistic regression and linear regression, respectively. We conducted separate models for each exposure variable with class membership as the dependent variable. Maximum likelihood adjustment was used to correct for classification error bias.

Results

General characteristics of the study sample

Data of 403 Mexican adolescent girls (mean age 15·2 (sd 1·8) years) were analysed, of whom 94 (23·3 %) had a 6-KADS score equal or higher to 6, indicating the evidence of depression. In total, 21·8 % of the participants suffered from anaemia and 10·9 % from iron deficiency. Overweight or obesity was present in almost half of the adolescents (42·7 %). Detailed descriptive statistics are shown in Table 1.

Table 1 Summary statistics of the total sample and the three latent classes of depressive symptoms

Depressive symptoms if 6-KADS score was ≥ 6; anaemia was defined as Hb levels < 120 g/l; iron deficiency was defined as concentrations of serum ferritin < 15 µg/l; iron-deficient erythropoiesis was defined as soluble transferrin receptor concentration > 8·3 mg/l. CRP, AGP: serum concentrations of C-reactive protein and α 1-acid glycoprotein.

* Mean values (sd).

Median values (Q25, Q75) or n (%).

Heterogeneity of depressive symptoms

The LCA analysis indicated three distinctive classes for the degree of depressive symptoms. When BMI-for-age was entered in the model, the Bayesian information criterion value changed from 4332·5 to 4321·3, indicating the most optimal balance between model fit and model simplicity, and was therefore the preferred solution for classification (Table 2). The percentages of individuals with self-reported scores of almost never in each of the six items of the 6-KADS questionnaire are shown in Table 3. Class 1 was the largest subgroup, constituting 44·4 % of the adolescent girls, who were ‘unlikely depressed’. Class 2 was labelled as ‘likely depressed’ because the response pattern reflected moderate occurrence (sometimes) for four out of six of the 6-KADS items. Class 3 was composed of 14·1 % of the adolescents and was described by a higher occurrence of depressive symptoms, with more adolescent girls reporting suicidal or self-harm ideation than the other two subgroups. Thus, the last class was labelled as ‘highly likely depressed’.

Table 2 Comparison of models with different number of classes derived from latent class analysis

Bayesian information criterion (BIC) and Akaike information criterion (AIC) are both indicators of parsimony, that is the balance between model fit and model simplicity; the lower their value, the better this balance. Number of parameters (Npar); the lowest Npar indicates the model with the most parsimony. Degrees of freedom (df). Likelihood ratio chi-squared goodness-of-fit statistic (L2) indicates the amount of the association among the variables that remained unexplained after estimating the model; the lower the value, the better the fit of the model to the data. P-value from the bootstrap provides a more precise estimate by relaxing the assumption that the L2 statistic follows a chi-square distribution; P-values < 0·05 indicate no improvement over the previous model.

* Model with BMI-for-age as continuous covariate with direct effect on 6-KADS questions 3 and 6. Bold values designate the selected model, with the best balance between model fit and model simplicity.

Table 3 Description of the selected 3-class latent class model

Values indicate conditional probabilities. For example, given the membership of class 1, the probability of responding ‘Almost never’ to the first item is 0·615. Entries are given as response probabilities in each item of the 6-KADS questionnaire. Low P-values indicate that the questionnaire item was useful to differentiate between class memberships.

Associations of class membership and iron status

We examined the association between the three depressive symptoms classes with iron deficiency and latent classes of depression, and iron-deficient erythropoiesis and latent classes of depression using multiple logistic regression models adjusted for other potential confounders, including years since menstruation, BMI z-scores and markers of inflammation (Table 4). Compared to their peers with normal ferritin concentrations, iron-deficient girls (serum ferritin < 15 µg/l) had higher odds of being ‘likely depressed’ and ‘highly likely depressed’ (OR 2·01; 95 % CI 1·01, 3·00 and OR 2·80; 95 % CI 1·76, 3·84, respectively). In contrast, we found no evidence of an association between depressive symptoms and iron-deficient erythropoiesis (soluble transferrin receptor concentration > 8·3 mg/l).

Table 4 Probability of latent class membership for exposure variables: results obtained by multinomial logistic regression analysis

sTfR, soluble transferrin receptor.

Values indicate OR, obtained by multinominal logistic regression analyses with class 1 (‘unlikely depressed’) as the reference category. P-values are based on Wald statistics. sTfR, CRP, AGP: serum concentrations of soluble transferrin receptor, C-reactive protein and α1-acid glycoprotein.

* Interpretation: girls with iron deficiency have 134·0 % higher odds of being ‘highly likely depressed’ compared with girls without iron deficiency.

Iron deficiency model adjusted for years since menstruation, BMI z-scores, CRP and AGP. Iron-deficient erythropoiesis model adjusted for the same variable than iron deficiency model and age at menarche.

Associations of class membership with Hb, body weight and puberty

The associations between depressive symptoms and Hb concentration, body weight and puberty are reported in Table 5. Each unit increment in Hb concentration (1 g/l) increased the probability of membership in class ‘unlikely depressed’ by 16 percentage points and reduced the probability of membership in class ‘likely depressed’ by 17 percentage points (P = 0·04). Each unit increment in BMI-for-age (1 sd) increased the probability of being ‘likely depressed’ by 22 % but reduced the probability of being ‘unlikely depressed’ by 18 % and ‘highly likely depressed’ by 4 %. Each increment in chronological age by 1 year increased the probability of being ‘highly likely depressed’ by 20 %. We assessed the association between puberty and depressive symptoms with two different indicators, years since menstruation and age at menarche, but we found no evidence of an association between puberty and depressive symptoms.

Table 5 Probability of latent class membership for exposure variables: results obtained by linear regression analysis

Values indicate β estimates with standard errors, obtained by multinominal linear regression analyses. P-values are based on Wald statistics.

* For example: probability of being member of the class ‘unlikely depressed’ = –1·65 + 0·16 (Hb, g/l). Interpretation: each increment in Hb, g/l by 1 unit increases the probability of membership in class ‘unlikely depressed’ by 16 percentage points.

Model adjusted for age at menarche and age.

Discussion

Three distinct subtypes of depression were derived from the LCA, in which the severity of the symptoms was the source of heterogeneity. Our results indicate that depression was associated with iron deficiency, low Hb concentration, higher body weight and age. There was no association between depression and puberty.

In the current study, iron-deficient girls had a higher chance to be ‘likely depressed’ (OR 2·01, 95 % CI 1·01, 3·00) or ‘highly likely depressed’ (OR 2·80, 95 % CI 1·76, 3·84) compared with non-iron-deficient girls. These findings are in accordance with other cross-sectional studies indicating that depression scores were higher in women with iron deficiency (OR 2·84 95 % CI 1·24, 6·51)(Reference Chen, Su and Chen26) and adolescents with iron deficiency anaemia (OR 2·34, 95 % CI 1·58, 3·46)(Reference Khoshgoo, Eslami and Khadem Al-Hosseini27). However, cross-sectional studies give no evidence of causality, and unfortunately, there is limited availability of longitudinal and intervention studies of iron status and depression among adolescents. A retrospective cohort study in Taiwan showed that 20-year-old women with iron deficiency anaemia had a 49 % increased risk of depression compared with women without iron deficiency anaemia(Reference Lee, Chao and Huang28). In addition, randomised controlled trials in women with postpartum depression demonstrated that iron supplementation improves the depressive symptoms in iron-deficient and non-iron-deficient women(Reference Wassef, Nguyen and St-André6). Previous studies have reported the effects of iron on different neurological activities, such as myelination and monoamine metabolism(Reference Kim and Wessling-Resnick29). Therefore, it is plausible that iron deficiency contributes to depression.

Another important finding was that a higher body weight was associated with being in the ‘likely depressed’ group and inversely associated with being ‘unlikely depressed’. Similarly, results from the National Survey of Health and Nutrition in Mexico (ENSANUT-2012) showed that adult women with obesity had higher odds (OR 1·28 95 % CI 1·07, 1·53) of having depression in comparison with normal-weight women(Reference Zavala, Kolovos and Chiarotto30). These findings are consistent with results from a meta-analysis that concluded that children and adolescents with obesity were more likely to be depressed (pooled OR 1·34; 95 % CI 1·1, 1·64), but no association was observed for their peers with overweight(Reference Quek, Tam and Zhang31). Some longitudinal studies have shown that depressive symptoms at baseline were associated with obesity after 1 year follow-up in white, black and Hispanic American adolescents(Reference Goodman and Whitaker32), in adults(Reference Tyrer, Zaccardi and Khunti33) and in women but not in men(Reference Vittengl34). In contrast, baseline obesity did not predict follow-up depression(Reference Goodman and Whitaker32). Whether the association between body weight and depression is bidirectional is not entirely elucidated as it is complex and is mediated by multiple biological pathways and psychosocial factors.

Contrary to expectations, this study did not find evidence of an association between depression and pubertal development. A possible explanation for the inconsistency with other studies is the use of different indicators for pubertal stage(Reference Keenan, Culbert and Grimm35Reference Carter, Silverman and Jaccard38). Puberty comprises two distinct but overlapping processes, adrenarche (early stage) and gonadarche (later stage)(Reference Mendle, Beltz and Carter39). A dramatic rise of steroid hormones marks adrenarche and is associated with physical changes that include increased skin oil and acne, skeletal maturation and pubic hair growth. Gonadarche instead is a gradual process that typically starts with breast development for girls and finalises shortly after menarche(Reference Mendle, Beltz and Carter39). Thus, pubertal timing and pubertal status play a different role in the development of depressive symptoms. Another possible explanation for this might be that the mean age at menarche was similar in the three subgroups of depression, indicating that most girls were in comparable pubertal stages. Future investigations should include participants at different pubertal stages and consider using various indicators of pubertal status and pubertal timing to elucidate the link between puberty and mental health.

The use of LCA as a statistical technique to identify subgroups of depression is a major strength of this study. The use of categorical diagnostic constructs can result in the loss of valuable information about the diagnosis, because those who score just below the diagnostic threshold are regarded as non-cases. LCA overcomes this pitfall and has additional benefits for understanding the biological pathways and treatment opportunities by discriminating subtypes of depression.

There were some limitations to this study and should be considered when interpreting our results. Primarily, because of the observational nature of our study, there may exist unmeasured or residual confounding in the associations of interest. Furthermore, while our analysis adjusted for common demographic variables, other psychosocial indicators, which are likely to encompass the biological pathways that influence mental health, were not measured. Despite these limitations, our results are a starting point for future studies to investigate the role of nutritional status on mental health.

In conclusion, iron-deficient adolescent girls were more likely to suffer from depressive symptoms, and low Hb concentration and higher body weight increased the probability of depression. These findings suggest clinical trials to determine if nutritional status plays a role in depression and if improving nutritional status may alleviate some symptoms of depression. In addition, greater focus on screening and detecting depression in adolescent girls, especially among those with poor nutritional status, is required.

Acknowledgements

Acknowledgements: The authors thank the principals and teachers of the schools for their support, as well as participating parents and the students for their cooperation. Financial support: This work was supported by Universidad de Monterrey, Mexico and The National Council of Science and Technology (CONACyT), Mexico [grant 460002]. Authorship: conceptualisation, A.G.Z.-O., A.M.-B., A.C.C.-L., A.Q.-G. and E.J.M.F.; methodology, H.V.; data collection A.G.Z.-O., B.-J. W. and E.E.L.-B.; formal analysis, A.G.Z.-O.; writing – original draft preparation, A.G.Z.-O.; writing – review and editing, A.C.C.-L., A.M.-B., H.V.,E.J.M.F., A.Q.-G., B.-J. W. and E.E.L.-B. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving research study participants were approved by the Committee of Ethics in Research (CIE) of the Health Science Division, University of Monterrey (Reference 04072018-CEI). Written and verbal informed consent was obtained from all participants and legal guardians.

Conflict of interest:

There are no conflicts of interest.

Supplementary material

For supplementary material referred to in this article, please visit https://doi.org/10.1017/S1368980022001203

References

Campisi, SC, Zasowski, C, Shah, S et al. (2020) Assessing the evidence of micronutrients on depression among children and adolescents: an evidence gap map. Adv Nutr 11, 908927.CrossRefGoogle ScholarPubMed
Georgieff, MK (2011) Long-term brain and behavioral consequences of early iron deficiency. Nutr Rev 69, S43S48.CrossRefGoogle ScholarPubMed
Gholamreza Noorazar, S, Ranjbar, F, Nemati, N et al. (2015) Relationship between severity of depression symptoms and iron deficiency anemia in women with major depressive disorder. J Anal Res Clin Med 3, 219224.CrossRefGoogle Scholar
Saluja, G, Iachan, R, Scheidt, PC et al. (2004) Prevalence of and risk factors for depressive symptoms among young adolescents. Arch Pediatr Adolesc Med 158, 760765.CrossRefGoogle ScholarPubMed
Webb, P, Nishida, C & Darnton-Hill, I (2007) Age and gender as factors in the distribution of global micronutrient deficiencies. Nutr Rev 65, 233245.CrossRefGoogle ScholarPubMed
Wassef, A, Nguyen, QD & St-André, M (2019) Anaemia and depletion of iron stores as risk factors for postpartum depression: a literature review. J Psychosom Obstet Gynecol 40, 1928.CrossRefGoogle ScholarPubMed
Oddy, WH, Allen, KL, Trapp, GSA et al. (2018) Dietary patterns, body mass index and inflammation: pathways to depression and mental health problems in adolescents. Brain Behav Immun 69, 428439.CrossRefGoogle ScholarPubMed
Chiang, JJ, Bower, JE, Irwin, MR et al. (2017) Adiposity moderates links from early adversity and depressive symptoms to inflammatory reactivity to acute stress during late adolescence. Brain Behav Immun 66, 146155.CrossRefGoogle ScholarPubMed
Ministry of Public Education (SEP) (2009) First National Survey of exclusion, intolerance and violence in public secondary schools. https://catedraunescodh.unam.mx/catedra/BibliotecaV2/Documentos/Educacion/Informes/SEP_Encuesta_Discriminacion_y_Violencia_2008.pdf (accessed December 2021).Google Scholar
Deecher, D, Andree, TH, Sloan, D et al. (2008) From menarche to menopause: exploring the underlying biology of depression in women experiencing hormonal changes. Psychoneuroendocrinology 33, 317.CrossRefGoogle ScholarPubMed
Colich, NL, Kircanski, K, Foland-Ross, LC et al. (2015) HPA-axis reactivity interacts with stage of pubertal development to predict the onset of depression. Psychoneuroendocrinology 55, 94101.CrossRefGoogle ScholarPubMed
Cepeda-Lopez, AC, Osendarp, SJM, Melse-Boonstra, A et al. (2011) Sharply higher rates of iron deficiency in obese Mexican women and children are predicted by obesity-related inflammation rather than by differences in dietary iron intake. Am J Clin Nutr 93, 975983.CrossRefGoogle ScholarPubMed
Vanes, LD, Moutoussis, M, Ziegler, G et al. (2020) White matter tract myelin maturation and its association with general psychopathology in adolescence and early adulthood. Hum Brain Mapp 41, 827839.CrossRefGoogle ScholarPubMed
Huberty, TJ (2012) The Developmental Psychopathology of Depression BT – Anxiety and Depression in Children and Adolescents: Assessment, Intervention, and Prevention. New York: Springer New York.CrossRefGoogle Scholar
Lee, Y-J & Kim, H-B (2020) Association between anaemia and adult depression: a systematic review and meta-analysis of observational studies. J Epidemiol Community Health 74, 565.Google ScholarPubMed
Shariatpanaahi, MV, Shariatpanaahi, ZV, Moshtaaghi, M et al. (2007) The relationship between depression and serum ferritin level. Eur J Clin Nutr 61, 532535.CrossRefGoogle Scholar
Ellis, R, Fernandes, A, Simmons, JG et al. (2019) Relationships between adrenarcheal hormones, hippocampal volumes and depressive symptoms in children. Psychoneuroendocrinology 104, 5563.CrossRefGoogle ScholarPubMed
LeBlanc, JC, Almudevar, A, Brooks, SJ et al. (2002) Screening for adolescent depression: comparison of the Kutcher adolescent depression scale with the Beck depression inventory. J Child Adolesc Psychopharmacol 12, 113126.CrossRefGoogle ScholarPubMed
Romero-Acosta, K, Gómez-de-Regil, L, Lowe, G et al. (2018) The relationship between substance use and depressive symptoms among Colombian adolescents. West Indian Med J 67, 323327.CrossRefGoogle Scholar
Erhardt, JG, Estes, JE, Pfeiffer, CM et al. (2004) Combined measurement of ferritin, soluble transferrin receptor, retinol binding protein, and C-reactive protein by an inexpensive, sensitive and simple sandwich enzyme-linked immunosorbent assay technique. J Nutr 134, 31273132.CrossRefGoogle ScholarPubMed
World Health Organization (2020) WHO Guideline on Use of Ferritin Concentrations to Assess Iron Status in Individuals and Populations. Geneva: World Health Organization.Google Scholar
World Health Organization (2011) Haemoglobin Concentrations for the Diagnosis of Anaemia and Assessment of Severity Vitamin and Mineral Nutrition Information System. Geneva: World Health Organization.Google Scholar
World Health Organization (2007) Growth Reference Data for 5–19 Years. Geneva: World Health Organization.Google Scholar
Kongsted, A & Nielsen, AM (2017) Latent class analysis in health research. J Physiother 63, 5558.CrossRefGoogle ScholarPubMed
Vermunt, JK & Magidson, J (2020) How to perform three-step latent class analysis in the presence of measurement non-invariance or differential item functioning. Struct Equ Model 28, 356364.CrossRefGoogle Scholar
Chen, M-H, Su, T-P, Chen, Y-S et al. (2013) Association between psychiatric disorders and iron deficiency anemia among children and adolescents: a nationwide population-based study. BMC Psychiatr 13, 161.CrossRefGoogle ScholarPubMed
Khoshgoo, M, Eslami, O, Khadem Al-Hosseini, M et al. (2020) The relationship between household food insecurity and depressive symptoms among pregnant women: a cross sectional study. Iran J Psychiatr 15, 126133.Google ScholarPubMed
Lee, H-S, Chao, H-H, Huang, W-T et al. (2020) Psychiatric disorders risk in patients with iron deficiency anemia and association with iron supplementation medications: a nationwide database analysis. BMC Psychiatr 20, 216.CrossRefGoogle ScholarPubMed
Kim, J & Wessling-Resnick, M (2014) Iron and mechanisms of emotional behavior. J Nutr Biochem 25, 11011107.CrossRefGoogle ScholarPubMed
Zavala, GA, Kolovos, S, Chiarotto, A et al. (2018) Association between obesity and depressive symptoms in Mexican population. Soc Psychiatr Psychiatr Epidemiol 53, 639646.CrossRefGoogle ScholarPubMed
Quek, YH, Tam, WWS, Zhang, MWB et al. (2017) Exploring the association between childhood and adolescent obesity and depression: a meta-analysis. Obes Rev 18, 742754.CrossRefGoogle ScholarPubMed
Goodman, E & Whitaker, RC (2002) A prospective study of the role of depression in the development and persistence of adolescent obesity. Pediatrics 110, 497504.CrossRefGoogle ScholarPubMed
Tyrer, F, Zaccardi, F, Khunti, K et al. (2020) Incidence of depression and first-line antidepressant therapy in people with obesity and depression in primary care. Obesity 28, 977984.CrossRefGoogle ScholarPubMed
Vittengl, JR (2018) Mediation of the bidirectional relations between obesity and depression among women. Psychiatr Res 264, 254259.CrossRefGoogle ScholarPubMed
Keenan, K, Culbert, KM, Grimm, KJ et al. (2014) Timing and tempo: exploring the complex association between pubertal development and depression in African American and European American girls. J Abnorm Psychol 123, 725736.CrossRefGoogle ScholarPubMed
Thompson, SM, Hammen, C & Brennan, PA (2016) The impact of asynchronous pubertal development on depressive symptoms in adolescence and emerging adulthood among females. J Youth Adolesc 45, 494504.CrossRefGoogle ScholarPubMed
McGuire, TC, McCormick, KC, Koch, MK et al. (2019) Pubertal maturation and trajectories of depression during early adolescence. Front Psychol 10, 1362.CrossRefGoogle ScholarPubMed
Carter, R, Silverman, WK & Jaccard, J (2013) Race and perceived pubertal transition effects on girls’ depressive symptoms and delinquent behaviors. J Youth Adolesc 42, 11551168.CrossRefGoogle ScholarPubMed
Mendle, J, Beltz, AM, Carter, R et al. (2019) Understanding puberty and its measurement: ideas for research in a new generation. J Res Adolesc 29, 8295.CrossRefGoogle ScholarPubMed
Figure 0

Table 1 Summary statistics of the total sample and the three latent classes of depressive symptoms

Figure 1

Table 2 Comparison of models with different number of classes derived from latent class analysis

Figure 2

Table 3 Description of the selected 3-class latent class model

Figure 3

Table 4 Probability of latent class membership for exposure variables: results obtained by multinomial logistic regression analysis

Figure 4

Table 5 Probability of latent class membership for exposure variables: results obtained by linear regression analysis

Supplementary material: File

Zarate-Ortiz et al. supplementary material

Zarate-Ortiz et al. supplementary material 1

Download Zarate-Ortiz et al. supplementary material(File)
File 21.4 KB
Supplementary material: PDF

Zarate-Ortiz et al. supplementary material

Zarate-Ortiz et al. supplementary material 2

Download Zarate-Ortiz et al. supplementary material(PDF)
PDF 33 KB