Sampling design
A sample of West Bank students were randomly selected from UNRWA and government schools for a 2009 cross-sectional school-based survey. The UNRWA and the government provide the bulk of free education in Palestine, serving over 85% of students [16,17]. In order to sample the schools, the West Bank was divided into three geographical areas (north, middle, and south) each accounting for roughly one third of the total school children population. Schools with at least 50 students (both girls and boys) in three selected grades were sampled for this study. UNRWA and government school students were stratified by region, sex, and school year (grade). The sample size required to report prevalence data (i.e., the percentage of first graders with stunting) was calculated using a one-sample proportion tests, with 80% power for a pre-set type 1 error of 5% for the main variables to be investigated. The proportion test returned a required sample size of 500 for each age group included in the study. The three age groups were as follows: first grade (mean age 6.71 years [SD 0.45]), sixth grade (mean age 11.82 years [SD 0.57]), and ninth grade (mean age 14.83 years [SD 0.61]).
The initial plan was to select six government schools and six UNRWA schools to cover the north, central, and southern regions of the West Bank, and to then randomly select 42 students from each grade and class. This would have yielded 504 (42*12) students from each grade, with 252 boys and 252 girls. The plan had to be amended to accommodate the realities of UNRWA and government schools in the region.
First, students were randomly recruited from 16 rather than 4 UNRWA schools. This resulted from the selection of two camps from each region and the fact that in some UNRWA schools, the elementary school and preparatory school are separate (requiring the selection of four schools rather than two). In light of these considerations, six government and 16 UNRWA schools were selected from 1555 UNRWA and government schools. Second, it was found that some of the government schools incorrectly reported the number of students in each class; in some cases, the school did not have enough students to allow the random selection of 42 students from each grade. Consequently, only 681 students were selected from government schools (rather than 756 as initially planned). From UNRWA schools 819 students were then selected to make up the required total of 1500. One student per household was selected to avoid clustering.
For each study index, two measurements of the standing height and weight (lightly dressed, shoes removed) were taken by trained field workers on visits to the participating schools. A third measurement was taken if the difference between the two measurements was greater than 10%. Mothers were invited to attend their children’s schools on the day of the visit and were asked to provide information regarding socio-demographic characteristics. In the case of first grade students, mothers were asked to complete the student survey on behalf of their children. Sixth and ninth grade students were asked to complete a survey regarding their physical activity and the number of hours spent watching TV, among other variables.
Laboratory measurements
Blood (5 mL) was obtained by venipuncture. To check validity of test results, two types of quality assurance program were used; internal and external quality control. Internal quality control was carried out for all lab tests, where three different levels of control were used with every batch of samples analyzed. Assays were externally validated for vitamin B12, ferritin, and folic acid levels using control quality assurance samples provided by the Centers for Disease Control and Prevention (CDC). The CDC - External Quality Assurance (EQA) program is a standardization program designed to provide laboratories with an independent assessment of their analytical performance. The blind samples were sent by the CDC-EQA program to Ramallah. A complete blood cell count was performed with a Fully Automated Hematology Analyzer by Cell Dyne 1700an. Thyroid hormones, vitamin B12, folate, and ferritin were determined using a Chemiluminesence assay using ADVIA Centaur USA, by Siemens. C-reactive protein (CRP) was determined by the ELISA technique using commercially available kits (Quantikine Human, R&D Systems, Minneapolis, MN). Zn content was measured using atomic absorption spectrophotometer blood. Lead was determined in a randomly selected 50% of the samples using Leadcare Analyzer, ESA Biosciences. A capillary tube was used to draw a small amount of blood (approximately two drops) and deposit it in the Leadcare system. Leadcare controls were done for each test kit.
In addition to the internal quality control procedures, we took the following measures in examining Fe, Zn, and lead:
- The venipuncture site was cleaned with alcohol.
- A closed-tube vacuum system was used to avoid mineral contamination.
- Special metal-free blood containers were used to minimize the potential for sample contamination by any outside sources of minerals.
- Blood was drawn in a BD Royal Blue with K2 EDTA Vacutainer tube (Supply T183).
- Lead blood specimens were drawn in a BD Tan with K2 EDTA, lead only (EDTA) Vacutainer tube (Supply T615).
- All tubes were kept in dark cool boxes (0-4°C) and transported to the central lab.
For quality control, double data entry was used for all lab tests, in addition to survey data. Students with elevated CRP were excluded from ferritin and zinc analyses.
Variables
Low Mean Corpuscular Volume (MCV) was defined as MCV<75fL
Iron deficiency was defined as serum ferritin below 15 ng/ml based on WHO guidelines [18].
Anemia: Based on WHO guidelines, anemia was defined as hemoglobin (Hb) below 11.5 g/dl for children 5-11 years; below 12 g/dl for children 12-14 years; and girls ≥15, and below 13 g/dl for boys ≥15 years [18].
Iron deficiency anemia: was defined as having both anemia and iron deficiency.
Serum folate: levels of 7-20 µg/L were defined as normal; levels of <3.1 µg/L were defined as very low, levels of <7.0 µg/L) were defined as low, and levels of >20 µg/L were defined as high [19].
Iodine deficiency: iodine deficiency was defined by a low thyroxine which was determined according to the manufacturer’s instructions as tri-iodothyroxine (FT3) <2.3 pg/ml, thyroxine hormone (FT4) <0.89 g/dl, and thyroid-stimulation hormone (TSH)>5.5 uIU/ml.
Vitamin B12: levels were defined as B12 deficient (<221 pg/ml) and marginal (<300 pg/ml) [20].
Serum Zinc: low levels were defined as <65 µg/dl [21].
C-reactive protein (CRP): Elevated levels were defined >11 mg/l [22].
Lead: High levels were defined as >10 µg/dl [23].
Statistical analysis
SAS (SAS Institute, Cary, NC) software was used to analyze the dataset of 1484 subjects. Means and percentages were used to describe the characteristics of the study sample. Chi-square test was performed to examine differences in micronutrient deficiency by gender, school affiliation, and grade level. Tests of significance were two-sided with p- value ≤0.05. The effect of fortification on the prevalence of micronutrient deficiencies was evaluated based on previous reports and an earlier baseline study in Hebron/West Bank in 2005 [24].