This study was carried out in the semi-rural communities of Muea, Ekona and Dibanda, situated at the foot of Mount Cameroon in the South West Region of Cameroon, as shown in Figure 1 . Muea is a semi-rural setting in the rain forest ecozone on the Eastern flank of the active volcanic Mount Cameroon, and it is located 562 m above sea level at longitudes 09˚18'53"E and latitudes 04˚05'14"N and has been described in detail by [10] . The rainy season spans from March to October, and the dry season, characterized by frequent light showers, runs from November to February.

Dibanda is located on the windward side of the eastern slope of Mount Cameroon with an altitude of 530 m above sea level, 25 km from the Atlantic Ocean, at longitude 09˚18'47"E and latitude 04˚05'09"N and has been described by [11] .

Ekona has two distinct seasons: a cold, rainy season from March to October, and a warm, dry season with frequent light showers from November to February. It is located 898 m above sea level at longitudes 9˚20'4"E and latitude 4˚14'0"N. The majority of inhabitants are involved in farming, trading or livestock keeping. The two seasons of Ekona have been described in detail by [12] .

All children ≤ 5 years of age resident in Muea, Ekona and Dibanda whose parents/caregivers consented to participate in the study were recruited. Children with severe malaria (unable to drink or breastfeed, vomiting more than twice in the proceeding 24 hrs before presentation, recent history of convulsion, unconscious state or unable to stand and other diseases requiring hospital admission); Severe complications of malnutrition requiring hospitalization in intensive care or stabilization, including kwashiorkor; Known underlying chronic or severe disease (e.g., HIV/AIDS, TB, cardiac, renal or hepatic disease, sickle cell); Presence of febrile conditions due to other diseases which could alter the outcome of the study were excluded from the study.

This observational cross-sectional study was carried out between the months of August 2016 and August 2018, including the rainy season (March-October). After obtaining ethical clearance and administrative and local authorizations from the chief and block heads in the neighbourhood, the study team proceeded to the field for sample collection.

The minimum number of samples required for the study was calculated using the formula,

$n=z_{\alpha /2}^2\left(1-a\right)P\left(1-P\right)/d^2$ (1)

where n is the minimum sample size required; z is 1.96, which is the standard normal deviate; a is absolute precision at 5%; P is 66.2% which is the proportion of anaemia prevalence [13] , and d is 0.05 (5%), the required margin of error. This gave a minimum sample size of 327 participants for each locality. Considering a possible participation of more than one child per family, loss of samples due to blood clotting and incomplete data entry, the sample size was adjusted by 10% to a minimum of 360.

A multi-stage sampling method was used in all the blocks in each neighbourhood in all three communities until the required sample was attained. At the start of the study in each site, the parents and guardians of the children were educated on the study protocol and the benefits of participation were highlighted at their various neighbourhoods. Ensuing administrative and ethical approval for the study, informed consent and assent forms explaining the purpose, risks, and benefits of the study were given to parent/caregivers. Participants were invited to the data collection location in each community by their local chiefs, and coordination was organised by the block heads of the various communities. Upon obtaining consent/assent from the participants, semi-structured questionnaires were administered to the mothers/caregivers to get socio-demographic data and information on the different feeding methods. Following administration of the questionnaire, anthropometric measurements were obtained and the blood sample was collected from each child for a full blood count assessment.

Questionnaire Survey

Data was collected regarding the following: demographics (gender and age), feeding habits (exclusive breastfeeding and duration/mixed feeding/no breastfeeding), types of local weaning foods, mother’s knowledge on balance diet, history of fever in the preceding 2 - 3 days, history of consumption of any medication in the preceding month, history of diarrhoea, availability of toileting facilities in household, sources of household water and parent/guardian educational level. Infants were classified as being exclusively breastfed using the definitions proposed by [14] . A child was said to be given mixed feeding (MF) when he/she was given a combination of breast milk and local infant formula before six months, and the last category of children were those who were not given breast milk (NBM) at all from birth and were fed with local infant formula.

Anthropometric measurements such as height and weight were measured using a measuring tape and a Terraillon weighing scale (Terraillon, Paris). The ages of the children were obtained from their mothers and caregivers. Under-nutrition indices such as height-for-age (HA), weight-for-age (WA), and weight-for-height (WH) standard deviation (SD) scores (Z scores) were computed based on the WHO growth reference curves using the WHO AnthroPlus for personal computers manual [4] . A child was identified as being malnourished if he or she scored <−2 in one of the anthropometric indices of HA (stunting), WA (underweight) and WH (wasting) indices, while corresponding Z scores of <−3 SD were considered indicative of severe under-nutrition [15] .

Four millilitres of venous blood samples were collected using sterile disposable syringes from children whose parents/caregivers signed the consent/assent forms. The collected blood was aliquoted into labelled ethylenediaminetetraacetate (EDTA) tubes. Labelled blood samples in EDTA tubes were transported on ice in a cool box to the Malaria Research Laboratory, University of Buea, for a full blood count analysis.

Haematological parameters were assessed using an auto-haematology analyser (MINRAY 2800 BC), following the manufacturer’s instructions. A complete blood count was obtained for haematological parameters such as haemoglobin levels, red blood cell counts (RBC), erythrocyte indices (mean corpuscular volume (MCV), mean corpuscular haemoglobin (MCH), mean corpuscular haemoglobin concentration (MCHC) and red cell distribution width (RDW), total white blood cell count (WBC) and platelet count (PLT). Anaemia was defined as Hb < 11.0 g/dL [WHO 2001] and further categorized as severe (Hb < 7.0 g/dL), moderate (Hb between 7.0 and 10.0 g/dL), and mild (>10 Hb < 11 g/dL).

Statistical Analysis

Data was cleaned up and analysed using the IBM-Statistical Package for Social Sciences (IBM-SPSS) version 20. Continuous variables were summarized into means and standard deviations (SD), and categorical variables reported as frequencies and percentages were used to evaluate the descriptive statistics. The differences in proportions were evaluated using Pearson’s Chi-Square (χ²). Group means were compared using analysis of variance (ANOVA), Student’s t-test, or Kruskal-Wallis test where appropriate. Significant levels were measured at 95% confidence interval (CI) with significant differences set at P < 0.05.

The study was approved by the Institutional Review Board hosted by the Faculty of Health Sciences, University of Buea (2021/004/UB/FHS/IRB) following administrative clearance from the South West Regional Delegation of Public Health, Cameroon. Informed consent/assent forms were given or read and explained to parents or caregivers of the children at presentation. The purpose and benefits of the study, as well as the amount of blood to be collected from each child, were clearly stated on the consent/assent forms. Only participants who gave written and/or verbal consent or assent took part in the study. Participation was strictly voluntary, and parents or caregivers were free at any point in time to stop the participation of the child/children in the study. All cases of moderate to severe anaemia as well as malnutrition were referred to the nearest health centre for appropriate treatment and follow-up, and the only little discomfort the children had was the pain perceived during blood collection from the vein.

The socio-demographic and clinical characteristics of the study participants are shown in Table 1 . A total of 1227 children with a mean (SD) age of 3.1 (1.6) years, residing in Ekona (35.5%, 436), Dibanda (31.9%, 391) and Muea (32.6%, 400) in the Mount Cameroon area were evaluated. The females were more (52.2%) than the males (47.8%), although not significantly. Participation within the age group was not significantly different, even though more of the children were of the ≤2 years (34.9%) age group. A greater proportion of the parents/guardians of the children (37.1%) had no formal education.

The prevalence of anaemia and malnutrition in children ≤ 5 years was 77.3% and 32.6% respectively. The most common form of malnutrition was stunting (27.1%).

The overall proportion of children who had EBF, MF and NBF from the 1227 children was 22.6%, 60.1% and 17.3% respectively. As shown in Table 2 , the highest proportion of mothers who exclusively breast fed their children was from Muea (26.8%), followed by Dibanda (23.3%) and least from Ekona (18.1%). Significantly (P = 0.004), most parents in the three study sites gave their children mixed feeding from birth. Among the 1196 children whose parent gave information on their educational level, a statistically significant (P < 0.001) proportion of parents with tertiary education exclusively breastfed their children (34.0%) while, parents with no formal education had the highest proportion of children who were given mixed feeding (64.2%) and those not breast-fed (20.5%).

*Information obtained from 1196 parents/caregivers only.

Children who were not breastfed at all had similar prevalence of malnutrition (34.0%), with MF (32.8%) and exclusively breast-fed children (31.0%). Across the different forms of malnutrition, children who were not breastfed at all had the highest prevalence of wasting (9.4%) that approached significance (P = 0.058) and underweight (11.3%) when compared with their counterparts ( Figure 2 ).

The prevalence of malnutrition in EBF, MF and NBF children was comparable across the different age groups. However, the prevalence of malnutrition among MF children was highest in the ≤2 years age group (38.3%) and the difference was significant at P = 0.001, as shown in Table 3 .

**statistically significant at P < 0.01; ***statistically significant at P < 0.001.

On the other hand, the prevalence of malnutrition in EBF, MF and NBF children was similar in the different sexes. Also, the prevalence of malnutrition was similar in males and females among EBF, MF and NBF children.

The prevalence of anaemia was significantly higher (P < 0.001) in children who had MF (80.5%) when compared with EBF (67.5%) children. The prevalence of severe and moderate anaemia was significantly higher among NBF (6.6%, 67.1%) when compared to their counterparts at P = 0.029 as revealed in Table 4 .

*Statistically significant at P < 0.05; ***statistically significant at P < 0.001.

The effect of age and sex on the prevalence of anaemia stratified by feeding habit is represented in Table 5 . The prevalence of anaemia varied significantly with age across the different feeding methods (EBF: P <0.001; MF: P = 0.001; and NBF: P = 0.01). Children who were EBF had a significantly lower (P < 0.001) prevalence of anaemia when compared with those who had other feeding methods among the 4.1 - 5.0 years old children. However, the prevalence of anaemia across the different feeding methods was comparable among the ≤2 and 2.1 - 4.0 years children.

*Statistically significant at P < 0.05; ***statistically significant at P < 0.001.

The prevalence of anaemia varied significantly within sex across the different feeding habits, with the males (72.0%) and females (63.4%) who were EBF having significantly lower (P = 0.002) prevalence of anaemia. The prevalence of anaemia was significantly higher in males (85.4%) than in females among MF children (P = 0.001) only, when the sexes were compared ( Table 5 ).

The mean haematological values for WBC × 10⁹/L, RBC × 10¹²/L, Hb (g/dL), HCT (%), MCV (fL), MCH (pg), MCHC (g/L), RDW-CV (%), PLT (L) among the children in the study area was 8.2 (2.3), 4.1 (1.4), 9.9 (1.5), 26.8 (4.3), 67.7 (7.0), 24.8 (4.0), 36.8 (5.0), 13.1 (3.8), 253.2 (130.1) respectively. The mean haematological parameters were comparable for the different sexes, except for haemoglobin concentration and haematocrit levels. The mean haemoglobin concentration (9.7 g/dL) and haematocrit levels (26.3%) were significantly lower in males when compared to females (P < 0.001), as shown in Table 6 .

Overall, the mean haematological parameters were comparable between children EBF, MF and NBF except for the RBC × 10¹²/L counts, Hb levels (g/dL), haematocrit levels (%), MCH (pg) and RDW-CV (%) as shown in Table 7 . Children NBF had significantly lower mean RBC × 10¹²/L counts [3.9 (1.0)], HB levels (g/dL) [9.6 (1.5)], HCT levels [26.0 (4.5)] and platelet count × 10⁹/L (235.8 ± 119.5).

*Statistically significant at P < 0.05; **statistically significant at P < 0.01; ***statistically significant at P < 0.001.

The overall anaemia prevalence in the study population was 77.3%. Although not significant, the prevalence of anaemia was lowest in Muea. Practically, difference between Ekona and Dibanda was minimal. The prevalence of anaemia was significantly higher in males (80.7%) and children ≤ 2 years of age (86.7%) when compared with members of their respective clusters, as shown in Table 8 . Relating to the severity of anaemia, the prevalence of severe and moderate anaemia was highest in males (5.7%, 60.3%), children ≤ 2 years of age (6.2%, 61.7%) at P = 0.049 and P < 0.001 respectively ( Table 8 ).

***Statistically significant at P < 0.001.

As shown in Figure 3 , malnourished children had significantly lower mean haemoglobin levels when compared with well-nourished children (P = 0.027). In addition, stunted children had a significantly lower mean haemoglobin level when compared with non-stunted (P = 0.037). The mean haemoglobin level was lower among underweight and wasted children, although the difference was not significant (P = 0.443 and P = 0.431 respectively).

The logistic regression model with malnutrition as the dependent variable and age, gender, level of education and marital status of guardian, anaemic status as well as infant feeding methods as the independent variables revealed age as the only significant predictor. Children ≤ 2 years and those between 2.1 and 4.0 years were 1.73 and 1.59 times respectively, more likely to be undernourished than children between 4.1 and 5.0 years old, as shown in Table 9 .

P values in bold are statistically significant.