Factors associated with iron deficiency anaemia among pregnant teenagers in Ashanti Region, Ghana: A hospital-based prospective cohort study

Background Iron Deficiency Anaemia (IDA) is reportedly high in pregnant adults and the causes well studied. However, among pregnant teenagers, the levels and associated factors of IDA are not fully understood. Methods In a prospective cohort study among Ghanaian pregnant teenagers, aged 13–19 years, IDA prevalence and associated factors were investigated. Sociodemographic data, household hunger scale (HHS), lived poverty index (LPI), FAO’s women’s dietary diversity score (WDDS) and interventions received during antenatal care (ANC) were obtained from 416 pregnant teenagers in Ashanti Region, Ghana. Micronutrient intakes using a repeated 24-hour dietary recall, and mid-upper arm circumference (MUAC) were determined and blood samples analysed for haemoglobin (Hb), serum levels of ferritin, prealbumin, vitamin A, total antioxidant capacity (TAC), C-reactive protein (CRP), and zinc protoporphyrin (ZPP). Results Anaemia (Hb cutoff <11.0 g/dL) was 57.1%; deficient systemic supply of iron stores (31.4%), depleted body stores of iron (4.4%), inadequate dietary iron intake (94.5%), and inadequate multiple micronutrient intakes (49.5%), were all notable among study participants. Between-subject effects using Generalized Linear Modelling indicated malaria tablet given at ANC (p = 0.035), MUAC (p = 0.043), ZPP (p<0.001), ZPP/Hb ratio (p<0.001) and depleted body iron stores (DBIS) (p<0.001) to significantly affect Hb levels. Pregnant teenagers with a high ZPP/Hb ratio (OR = 9.7, p<0.001, 95%CI = 6.0–15.8) had increased odds of being anaemic compared to those with normal ZPP/Hb ratio. Participants who were wasted (OR = 1.2, p = 0.543, 95%CI = 0.6–2.3), and those with depleted iron stores (OR = 3.0, p = 0.167, 95%CI = 0.6–14.6) had increased odds of being anaemic. Participants who experienced hunger were close to 3 times more likely (OR = 2.9, p = 0.040, 95%CI = 1.1–7.8) for depleted iron stores, compared to those who did not experience hunger. Also, participants with inadequate multiple micronutrients intakes (OR = 2.6, p = 0.102, 95%CI = 0.8–8.4), and those with low serum levels of ferritin (OR = 3.3, p = 0.291, 95%CI = 0.4–29.2) had increased odds of depleted body iron stores. Conclusions IDA is common among pregnant teenagers and the related factors include malaria tablets given at ANC, maternal hunger, maternal MUAC, a deficient systemic supply of iron, depleted body iron stores, ZPP, and ZPP/Hb ratio. Appropriate interventions are urgently needed to address the causes of IDA among pregnant teenagers.

We used the first 24-hour recall from each participant to score her dietary diversity using the 10 food 163 groups. The WDDS was then classified as low (≤ 3), medium (4 -5 food groups), or high (≥ 6 food groups) 164 according to FAO's guidelines [41]. The WDDS was further regrouped into adequate (5 -9) and inadequate 165 (0 -4) dietary diversity scores to help with statistical association for the population. A questionnaire was 166 used to obtain data on food cravings, pica practices, and food aversions. 167 Household Hunger and Poverty status Assessment determined as a proxy indicator of maternal weight status since it has good specificity in determining 185 weight during pregnancy [45,46]. An inelastic tape measure was used to take MUAC of participants by 186 locating the midpoint between the acromion and olecranon bone on the left hand of participants. MUAC 187 measurement less than 22.0cm was referred to as severe wasting, 22.0cm to less than 24.0cm as 188 mild/moderate wasting, and 24.0cm and above as normal MUAC [47]. All anthropometric measurements 189 were done in duplicates and averages were used. 190 Determination of haemoglobin (Hb) levels 191 Two milliliters venous blood sample of participants collected into anticoagulant EDTA tubes was used to 192 determine haemoglobin levels of participants using Sysmex Haematology system (USA). The WHO cut-off 193 for Hb was used to determine anaemia prevalence among the participants; haemoglobin values less than 194 11.0g/dL means anaemia, 11.0g/dL and above means no anaemia [15]. Deficient systemic iron supply was defined as ZPP/Hb ratio greater than 80 μmol/mol heme, ZPP/Hb ratio 215 between 60-80 μmol/mol heme was termed moderate-high, and less than 60 μmol/mol heme means 216 adequate systemic iron supply [55]. Depleted body iron stores (DBIS) were defined as participants with 217 ZPP/Hb ratio greater than 80 μmol/mol heme, and serum ferritin less than 20μg/L. 218  Table 1 presents the relationship between iron deficiency status, and household hunger scale, lived 257 poverty, eating behaviours, and intervention received at ANC. Over 8 in 10 pregnant teenagers with 258 depleted body iron stores had inadequate women's dietary diversity score (WDDS) compared with less 259 than 6 in 10 (57.4%, p = 0.046) among those with normal body iron stores. Also, about 30% of pregnant 260 teenagers who were depleted of body iron stores suffered severe hunger compared with 11% of those 261 13 with normal body iron stores (p = 0.029). However, the proportion of participants who had depleted body 262 iron stores and those with normal body iron stores did not significantly vary by poverty status (p = 0.633), 263 food aversion (p = 0.133), food craving (p = 0.797), and whether the following services were received 264 during ANC or not: nutrition education (p = 0.808), micronutrient supplementation (p = 0.625), and 265 malaria tablet given (p = 0.315). 266

Data analysis
The proportion of participants who were anaemic (low Hb) and non-anaemic did not significantly differ 267   Table 2 presents the relationship between anthropometric status and anaemia (low Hb). The mean MUAC 314 was higher among pregnant teenagers who were not anaemic (26.6±0.2cm) than those who were anaemic 315 (25.8±0.2cm, p = 0.006) respectively. The proportion of participants who were anaemic and non-anaemic 316 did not significantly differ by BMI status (p = 0.417), and MUAC status (p = 0.319), although underweight 317 (2.6%), and severe wasting (6.5%) were both higher among anaemic participants than non anaemic 318 (underweight 1.1%, severe wasting 4.6%) participants. 319 The proportion of participants who had depleted body iron and those with normal body iron stores did 320 not significantly vary by BMI status (p = 0.147), and MUAC status (p = 0.454). 321 The relationship between serum nutrients status and iron deficiency anaemia (low Hb and depleted body 327 iron stores) status is presented in Table 3. The proportion of participants with a high ZPP/Hb ratio, 328 indicating poor systemic iron supply was higher in anaemic (49.3%) compared with non anaemic (7.7%) 329 participants (p<0.001). The mean levels of serum ZPP (p<0.001), and ZPP/Hb ratio (p<0.001) were 330 significantly higher among pregnant teenagers who were anaemic (38.3±2.9µmol/L, 109.6±9.0) than 331 those who were not anaemic (28.4±2.0µmol/L, 64.9±4.6) respectively. The proportion of anaemic and 332 18 non-anaemic participants with low serum vitamin A (p = 0.373), serum ferritin (p = 0.785), serum 333 prealbumin (p = 0.763) did not significantly differ. Also, the mean levels of serum c-reactive protein (p = 334 0.066) and total antioxidant capacity (p = 0.860) did not significantly differ between anaemic and non-335 anaemic participants. For serum iron status, more anaemic participants (6.8%) had depleted iron stores 336 than those who were not anaemic (1.2%, p = 0.010). 337 Almost all participants with deficient systemic iron supply (high ZPP/Hb ratio) (94.1%) had depleted body 338 iron stores compared with less than a third (28.5%) among those with normal body iron stores (p<0.001).  than the normal body iron stores participants, but highest among high ZPP/Hb ratio and also depleted 380 body iron stores participants. Last but not the least, anaemia was highest among low serum ferritin-381 depleted body iron stores participants (100%), followed by normal serum ferritin-depleted body iron 382 stores (87.5%), then low serum ferritin-normal body iron stores (61.5%) and the least were among normal 383 ferritin-normal body iron stores (55.2%) participants (p=0.010). 384   385  386  387  388  389  390  391  392  393  394  395  396  397  398  399  400  401  402   21   403  404  405  406  Table 4 The results of between-subject effects, conducted in the univariate Generalized Linear Model are 413 presented in Table 5. Taking into account all the multiple underlying factors of anaemia in the tests of 414 associations on haemoglobin (Hb) levels, it was malaria tablet given at antenatal clinic (p = 0.035), MUAC 415 (p=0.043), ZPP (p<0.001), ZPP/Hb ratio (p<0.001) and body iron stores (p<0.001) that had a significant 416 effect on the Hb levels. Some variables (e.gs. lived poverty index, serum levels of vitamin A, C-reactive 417 protein, prealbumin) which were not significant were removed, to report a few items in the table. 418 The predictors of anaemia and depleted body iron stores of the pregnant teenagers are presented in Table  425 6. Pregnant teenagers who had a high ZPP/Hb ratio (signifying poor systemic iron supply) (OR= 9.7, 426 23 p<0.001, 95%CI= 6.0-15.8) had more than 9 times increased odds of being anaemic compared to those 427 who had normal ZPP/Hb ratio. Pregnant teenagers who were wasted (OR= 1.2, p = 0.543, 95%CI= 0.6-2.3), 428 those who had depleted iron stores (OR= 3.0, p = 0.167, 95%CI= 0.6-14.6) had an increased odds of being 429 anaemic, compared to those with normal MUAC or normal iron stores respectively. 430 Also, pregnant teenagers who experienced hunger (OR= 2.9, p = 0.040, 95%CI= 1.1-7.8) had over 2 times 431 increased odds of having depleted body iron stores, compared to those who did not experience hunger.

453
The key findings of the study were the high prevalence of iron deficiency anaemia among teenagers. Iron 454 deficiency anaemia (IDA) was common among the pregnant teenagers, from both systemic serum iron, 455 and their diet, and this is worrying. Our finding on IDA among teenagers is higher than that of pregnant 456 adults in other studies in Ghana [20,56,57]. The studies by Ayensu et al. [57], the Ghana Demographic 457 Health Survey [56], and the Ghana Micronutrients Survey [20] have reported that anaemia prevalence 458 among pregnant adults was 56.0%, 44.6%, and 42.0% respectively. This means that compared to the 459 adults, teenagers are more likely to be anaemic during pregnancy due to their high nutritional needs and 460 immature body tissue growth. Higher anaemia among teenagers will more likely expose them to a greater 461 risk of poor pregnancy outcomes (e.g. low birth weight, small-for-gestation age, stillbirth, preterm birth) 462 than pregnant adults [16]. Almost 6 in 10 pregnant teenagers (57.1%) using Hb levels were anaemic, close 463 to a third (31.4%) had deficient systemic supply of body iron stores, 4.4% had depleted body stores of 464 iron, and almost all participants (97.6%) had inadequate dietary iron intake. Also, combined dietary 465 inadequacies for iron-folate (90.4%) and iron-vitamin B12 intakes (54.6%), and multiple micronutrient 466 deficiencies (52.2%) were very high, ranging from 5 to 9 in 10 participants. Iron deficiency anaemia is a 467 public health problem among pregnant teenagers in Ghana. Studies conducted in Palestine and Ghana 468 have reported a high prevalence of anaemia (25.0%, 70.0%, 52.0%), and depleted body iron stores (52.0%) 469 among pregnant teenagers/adults [21,32,58]. Anaemia in pregnant teenagers is attributed to the fact 470 that they are still growing and additional requirement for iron and folic acid to meet their own nutritional 471 needs as well as those of the developing foetus during gravidity [16] but the very low intakes of nutrients 472 that are linked with anaemia are equally attributable from this study. 473 Several studies have also reported sociodemographic status to affect iron deficiency anaemia in 474 pregnancy [9,32,59]. In this study, we found significant associations between sociodemographic 475 parameters of participants and iron deficiency anaemia. However, by comparison, the levels of anaemia 476 among rural participants, older teenagers (16-19 years), unmarried teenagers, and employed participants 477 were higher anaemic. Our findings are consistent with other studies that found no significant association 478 between monthly income level, level of education, household size, and anaemia prevalence [58, 60] but 479 26 also contrary to other studies that reported certain socio-demographic characteristics, such as being a 480 housewife, having less education, and multiparity to be associated with anaemia among pregnant women 481 [9,32,59]. 482 Food deprivation is associated with low systemic iron stores, and consequently anaemia [25]. For the 483 adolescent, pregnancy can be a complicated issue due to the rapid increased demands of nutrients, 484 especially for iron and folic acid, for the mother's blood expansion, growth as well as the foetus. Food 485 deprivation and poor access to food are linked with poor dietary intakes, and the risk of micronutrient 486 deficiencies, which can consequently lead to depleted body stores of nutrients [24, 27, 61]. We found 487 hunger to predict depleted iron stores, with teenagers who suffered hunger (OR= 2.9, p = 0.040, 95%CI= 488 1.1-7.8) compared to those who did not suffer from hunger at increased risk of depleted body iron stores. 489 We also found participants who were anaemic were more likely to suffer from both severe hunger and 490 depleted body iron stores, suggesting a combined effect. This implies that severe maternal hunger could 491 have caused depleted body iron stores, leading to the risk of anaemia. Although we observed that malaria 492 tablets received at ANC effectively determined haemoglobin levels of the participant, those who were 493 given tended to be more anaemic, implying that that malaria was not the likely cause of iron deficiency 494 anaemia among the teenagers. 495 The body requires iron, vitamin B12, and folate for the effective production of haemoglobin. Any deficiency 496 in these micronutrients are likely to cause anaemia due to inadequate production of red blood cells. Poor 497 dietary intake from inadequacies of micronutrients and poor dietary diversity are the main culprit for low 498 vitamin and mineral levels, and therefore anaemia [25,63]. We observed that teenagers who had 499 inadequate women's dietary diversity score (WDDS) compared to those with adequate WDDS were more 500 likely to have a depleted body iron stores (p= 0.046), and that, inadequate WDDS among the teenagers 501 significantly contributed to poorer iron stores and anaemia occurrence, but the regression model showed 502 a lower risk for depleted body iron stores (OR= 0.3, p = 0.089, 95%CI= 0.1-1.2). We also found that more 503 27 pregnant teenagers with depleted body iron stores had inadequate multiple micronutrient intakes than 504 those with normal body iron stores (p = 0.046), and that, inadequate multiple micronutrient intakes (iron-505 folate-vitamin B12) contributed to depleted body iron stores and anaemia occurrence in the teenagers. 506 This implies that consumption of less diversified diet together with multiple micronutrient inadequacies 507 during pregnancy is associated with depleted iron stores, and can consequently lead to maternal anaemia. 508 A study by Diana et al. [64] in Indonesia reported the consumption of a less diversified diet among 509 pregnant, which was not associated with iron inadequacy, contributed to anaemia in pregnant women. 510 Hence, supplementation of iron and folic acid to pregnant women need to be re-evaluated to factor the 511 additional demands of adolescent girls, and those who might be at risk of iron deficiency anaemia. Also, 512 appropriate nutrition education at the antenatal care clinics should focus on educating pregnant women 513 to consume nutritious and diversified diets, to improve the nutrition status of iron. 514 Most often, women in low-and-middle-income countries go into pregnancy malnourished, and the 515 demands of gravidity can worsen micronutrient deficiencies, leading to iron deficiency anaemia. In Ghana,516 Nonterah et al.
[32] studying both pregnant teenagers and adults reported multiparity and maternal 517 underweight as predictors of anaemia. In this study maternal BMI was not associated with iron deficiency 518 anaemia, although underweight prevalence was higher among anaemic teenagers compared to non-519 anaemic participants. On the other hand, the regression analysis showed that participants who were 520 wasted (using MUAC) (OR= 1.2, p = 0.543, 95%CI= 0.6-2.3) were more likely to be anaemic, and this was 521 consistent with a study among pregnant adults in Ghana [26]. 522 We observed that being depleted of body iron stores compared to normal body iron stores was associated 523 with an increased risk of anaemia. Deficient systemic supply of iron ( high ZPP/Hb ratio; (OR= 9.7, p<0.001, 524 95%CI= 6.0-15.8),) and depleted body iron stores (OR= 3.0, p = 0.167, 95%CI= 0.6-14.6) during pregnancy 525 were both predictors of iron deficiency anaemia in the teenagers, with an odds ratio greater than 2. 526 Participants with a deficient supply of iron were more likely to have poorer iron stores and become 527 28 anaemic. Similarly, participants with low serum ferritin levels were more likely to have poorer iron stores 528 and become anaemic. These imply that low serum ferritin and deficient systemic supply of iron could also 529 be associated with depleted iron stores in the teenagers, thus lead to the occurrence of maternal anaemia. 530 A depletion of iron stores occurs when there is a poor dietary intake, which affects the systemic supply of 531 iron for haemoglobin production [25,62]. In the body, ZPP and serum ferritin are among iron biomarkers 532 that affect iron status, and the combination of the two can predict whether iron stores are depleted or 533 not [55]. Ferritin is the main body storage of iron, and it can be elevated (above 200 µg/L) during 534 inflammation state, infection, which causes iron overload [49,55]. Wirth et al. [9] reported maternal 535 inflammation, vitamin A insufficiency, but not malaria, vitamin B12, and folate deficiencies to significantly 536 determine maternal anaemia. In this study, ferritin levels were less than 200µg/L for the population, which 537 depicts no likely inflammation or infection. Where there is no inflammation, the occurrence of low serum 538 ferritin, and deficient ZPP/Hb ratio depict poor dietary intakes and poor iron bioavailability [49]. Also, the 539 final step of haemoglobin production involves a combination of ferrous protoporphyrin and globin, but 540 where iron is deficient, iron is replaced by zinc to produce zinc protoporphyrin (ZPP) in a normal ratio 541 30,000:1. But progressive iron deficiency increases ZPP levels, which has a profound effect on the supply 542 of iron for haemoglobin production [48,66]. In this study, we observed a high ZPP/Hb ratio among the 543 teenagers who were anaemic, indicating deficient iron supply for haemoglobin production. All these 544 associations tend to explain that poor micronutrient intake and severe hunger, and low serum ferritin 545 could contribute to deficient systemic iron supply and concurrent depleted body iron stores, leading to 546 iron deficiency anaemia among pregnant teenagers. 547 The implication of this study is; in our previous study, we reported that a higher prevalence of anemia 548 among pregnant teenagers than adults, and mean Hb levels of both teenagers and adults were found to 549 be higher by the 36th week of pregnancy than at antenatal registration. Also, a positive correlation was 550 observed between Hb at antenatal registration and 36th-week gestation, indicating that pre-pregnancy 551 29 Hb levels, largely determined anaemia occurrence during the latter stages of pregnancy [21]. A follow-up 552 study was needed to understand the associated factors causing high anaemia among teenagers than 553 adults. And since most studies in Ghana had focused on pregnant adults, the teenagers who were at 554 greater risk of anaemia were considered for the current study. The current findings have provided enough 555 evidence that the high maternal anaemia prevalence among teenagers was a result of hunger experienced 556 during pregnancy, inadequate multiple micronutrient intakes, low serum ferritin levels which contributed 557 to the deficient systemic supply of iron and depleted body iron stores. Other factors such as wasting and 558 inadequate dietary diversity contributed to depleted iron stores, and anaemia occurrence. Additionally, 559 malaria tablets given at antenatal clinic, MUAC, ZPP, ZPP/Hb ratio, and body iron stores (p<0.001) had a 560 significant effect on the Hb levels of teenagers. Malaria tablet given at ANC affected Hb levels but was not 561 a significant contributor of iron deficiency anaemia in the teenagers. There is therefore the need for 562 nutrition education and women empowerment among these teenagers, and also antenatal service 563 delivery should comprehensively focus on the nutritional needs of pregnant women, while special 564 attention should be given to teenagers to prevent the devastating consequences of anaemia during 565 pregnancy. Also, interventions that encourage food fortification and diversification has been reported to 566 reduce anaemia in women of child-bearing age, and these interventions can be considered [67]. 567

568
Iron deficiency anaemia was found among pregnant teenagers in this study. Iron deficiency anaemia is 569 common among pregnant teenagers and the related factors include malaria tablets given at ANC, 570 maternal hunger, maternal MUAC, a deficient systemic supply of iron, and depleted body iron stores. 571 Antenatal service delivery should focus on the nutritional needs of pregnant women, while special 572 attention is given to teenagers to prevent the devastating consequences of anaemia during pregnancy. 573