Reference intervals for selected haematological and biochemical parameters among apparently healthy adults in different eco-geographical zones in Ghana

Background Due to the influence of gender, race/genetics, age, lifestyle habits and geography on the references intervals (RIs), the Clinical and Laboratory Standards Institute (CLSI) recommends the determination of population-specific RIs. Ghana continues to depend on pre-established RIs from other countries which poses the risk of misdiagnoses and wrong treatment. This study presents the haemato-biochemical RIs from four eco-geographical zones in Ghana. Methods In this population-based cross-sectional study, a total of 1227 randomly selected healthy voluntary blood donors from the four eco-geographic zones (Coastal Savannah, Rain Forest, Savannah and Transitional) were enrolled and screened. Based on the CLSI Guidance Document C28A2992, the data of eligible participants were used to non-parametrically determine the RIs for the haemato-biochemical parameters at the 2.5th and 97.5th percentiles. Comparison of analytes by gender was done by Wilcoxon rank sum test and eco-geographic differences were assessed using the Kruskal-Wallis with the Dunn post hoc multiple comparison tests. Results There were statistically significant differences in most of the haematological parameters (RBC, Hb, HCT, MCV, PLT, WBC; p-values <0.0001 and MCH; p-value = 0.007), and biochemical analytes (Urea, Cr, Trig, HDL-C, AST, ALT, ALP, GGT, BID, BIT, Prot-T and Albumin; p-values <0.0001) based on gender. Significant inter eco-geographic (intra-population) variations and substantial differences between the established RI and the RIs accompanying the analyzers used were also observed. Conclusion This study reports significant inter-sex and inter-geographical differences in haemato-biochemical RIs in Ghana as well as differences in RIs with both the RIs accompanying the analyzers and those of other countries. Determining RIs representative of populations and including them in the report systems of laboratories to ensure effective and efficient healthcare service delivery is thus recommended.


Methods
In this population-based cross-sectional study, a total of 1227 randomly selected healthy voluntary blood donors from the four eco-geographic zones (Coastal Savannah, Rain Forest, Savannah and Transitional) were enrolled and screened. Based on the CLSI Guidance Document C28A2992, the data of eligible participants were used to non-parametrically determine the RIs for the haemato-biochemical parameters at the 2.5 th and 97.5 th percentiles. Comparison of analytes by gender was done by Wilcoxon rank sum test and eco-geographic differences were assessed using the Kruskal-Wallis with the Dunn post hoc multiple comparison tests.

Results
There were statistically significant differences in most of the haematological parameters (RBC, Hb, HCT, MCV, PLT, WBC; p-values <0.0001 and MCH; p-value = 0.007), and biochemical analytes (Urea, Cr, Trig, HDL-C, AST, ALT, ALP, GGT, BID, BIT, Prot-T and Albumin; p-values <0.0001) based on gender. Significant inter eco-geographic (intra-population) variations and substantial differences between the established RI and the RIs accompanying the analyzers used were also observed. Introduction four zones, namely Coastal Savannah, Rain Forest, Transitional and Savannah zones. The Coastal Savannah zone is about 12,732 km 2 and covers predominantly Greater Accra and parts of Central and Volta regions, with average elevation between 20m and 150m above sea level. It has an annual mean temperature of 26˚C-30˚C, and mean rainfall of 750mm-850mm. The Rain Forest zone is about 61,651 km 2 and covers the Western, parts of Central, Eastern, Ashanti and Brong Ahafo regions, with average elevation between 15m and 240m above sea level. It has thick vegetation which contributes to a high amount of rainfall (�2000mm annually), rich soil types suitable for a wide range of crops cultivation and undulating lowlands with inselbergs. The Transitional zone has an area of about 65,152 km 2 and covers parts of Eastern, Ashanti, Volta and Brong Ahafo regions and has an average elevation between 150m and 640m above sea level. It has an annual mean rainfall of about 1250mm-1750mm and mean temperature of 26˚C-30˚C. The climatic features and vegetation makes this zone more suitable for agriculture [14,15]. The Savannah zone covers an area of about 99,182 km 2 , comprising Northern, Upper East and Upper West regions. It experiences one rainfall season annually (1,000mm-1,125mm) and the highest annual mean temperature of about 27˚C-36˚C. This zone has an average elevation between 180m and 300m above sea level and is the most sparingly populated and most of the populace practice subsistence farming [16]. Blood bank facilities within the four zones (37 Military Hospital-Accra (Coastal Savannah), Effia Nkwanta Regional Hospital-Sekondi/Takoradi (Rain Forest), Kwahu Government Hospital-Kwahu (Transitional) and Tamale Central Hospital-Tamale (Savannah)) were randomly selected as sampling sites.

Reference population, enrolment and sample size
This study involved randomly selected healthy adult voluntary blood donors between 18 to 59 years old. Eligible participants were interviewed via a general health status questionnaire (adapted from the CLSI Guidance Document C28A2) [17]. Pregnant and breastfeeding mothers, obesity (BMI >29 kg/m 2 ), evidence of medication use, use or abuse of alcohol and tobacco, presence of acute/chronic disease conditions, history of blood donation or transfusion within the last 3 months, surgery or hospitalization within the last 1 to 6 months, incomplete laboratory analysis results and any other confounding factors that may compromise the assessment of the analytes of interest were excluded from the study. According to the CLSI Guidance Document C28A2, a minimum of 120 participants are required in each group for non-parametric determination of RIs [17]. However, in an effort to improve the statistical power of the study, a total of 1227 healthy adult blood donors (at least 300 from each of the four eco-geographic zones) were invited to participate. Of the 1227 participants who were interviewed and screened, 235 were excluded due to possible confounding factors as shown in Fig 1. A total of 992 healthy blood donors were thus included in the final analyses. Biochemistry results were available for 874 participants. For the establishment of RIs based on gender to account for the differences in body physiology based on sex, males:500/433 and females: 492/431 were recruited for haematological/ biochemical RIs, respectively. For RIs based on co-geographical zones to account for the differences in geographical location, Coastal Savannah: 333/245, Rain Forest: 323/262, Savannah: 279/127, Transitional: 292/240 were recruited.

Sample collection and laboratory assays
About 8ml of venous blood sample was collected from each participant in the morning between 8:00 AM and 11:00 AM. About 5ml of the sample was dispensed into EDTA tubes (Jactermac, Germany) and the remaining 3ml was dispensed into vacuum gel tubes (Jactermac, Germany) for haematology and biochemistry assays, respectively. The needle of the syringe was removed prior to dispensing the blood into tubes to avoid hemolysis. The samples were transported in a cold box to the laboratory within 3 hours for laboratory analyses. The samples in the gel tubes were spun at 4000rpm for 10 minutes to obtain the serum. Screening for confounding factors were performed and cases such as G6PD deficiency, asymptomatic or sub-clinical malaria, Hepatitis B and C, HIV, sickle cell and other abnormal haemoglobin variants were excluded (Fig 1).

Confounding factors assessments
Sickle cell screening for each participant was done using 2% sodium metabisulphite and haemoglobin phenotype was investigated using alkaline electrophoresis at pH of 8.6. The methaemoglobin reductase technique was used for G6PD screening. Screening for syphilis (InTec-One-step immunoassay; Xiamen Inc., China), hepatitis B and C (Biotech Co. Ltd; Guangzhou Wondfo; China) and HIV (first response HIV1&2 card test; Premier Medical Corporation Ltd-India and oral quick by Orasure Technologies, Inc, Bethlehem, PA, USA) were done by rapid diagnostic test kits. Malaria was investigated using both rapid diagnostic test kits (CareS-tartTM -ACCESSBIO, USA) and 10% Giemsa-stained thick film for microscopy

Data management and statistical analysis
The study questionnaire was programmed into KoBoToolbox (https://www.kobotoolbox.org/ ), a web-based data collection software for ease of data collection and protection. Data was exported from the KoBoToolbox into Microsoft Excel Spreadsheet and verified. Stata version 14.1 (Stata Corp, College Park, TX, USA) was used for statistical analysis. The data was grouped based on gender and eco-geographical zones. The RIs were non-parametrically determined at 2.5th and 97.5th percentiles in accordance with the CLSI Guidance Document C28A2. Outliers estimated using the absolute difference between the most extreme distribution and the next value (D) and the Range (maximum-minimum) (R). Outliers were retained when D/R<0.33 [8]. Comparison of haemato-biochemical parameters by gender was performed by Wilcoxon rank sum test. Kruskal-Wallis tests with the Dunn post hoc tests were performed for differences between the four eco-geographical zones. A p-value of <0.05 was considered statistically significant.

Sociodemographic characteristics of the study population
The 992 participants comprised 245 (24.7%) from the Coastal Savannah zone, 262 (26.41%) from the Rain Forest zone, 245 (24.7%) from the Savannah zone and 240 (24.19%) from the Transitional zone. The mean age of the entire study population was 27.0 ± 6.7 years old. Other sociodemographic characteristics of study population stratified by eco-geographic zones are shown in Table 1.

Haematology reference intervals by gender
RBC count, Hb level, HCT, MCV, MCH, percentage and absolute eosinophil counts were significantly higher in males than females (p<0.05). However, RDW-CV, platelets and WBC were significantly higher in females than in males (p<0.0001) ( Table 2 and S1 Table).

Haematology reference intervals by eco-geographical zones
The Coastal Savannah zone presented with lower RBC count, Hb level, HCT and RDW-CV but higher absolute neutrophils among both males and females whereas the Transitional zone recorded the highest RBC count and MCHC. The Savannah had a higher RBC count relative to the Coastal Savannah and Rain Forest zones and a higher MON#, BAS# and BAS% compared to the Rain Forest zone. Hb level was higher in the Savannah zone compared to the Coastal Savannah zone. The Rain Forest recorded a higher RBC count and Hb level compared to the Coastal Savannah zone, and a higher MCV compared to the Savannah and Transitional zones. (Table 3 and S2 Table).

Biochemistry reference intervals by gender
Analytes of the liver function (AST, ALT, ALP, GGT, BID, BIT, Alb and Prot-T), serum urea, creatinine, Trig and LDL-C were significantly higher in males than in females (p<0.0001). Females however had a significantly higher HDL-C compared to males (p<0.0001) ( Table 4 and S3 Table).

Biochemistry reference intervals by eco-geographical zones
Coastal Savannah zone had higher BID, TChol and Trig but lower ALT, BIT, Albumin and Prot-T compared to Rain Forest zone. Savannah zone recorded a higher ALT, Albumin, Prot-T and HDL-C but lower TChol, Trig, LDL-C and urea compared to Coastal Savannah zone. Transitional zone presented with higher ALT, BIT, Trig and creatinine compared to the Rain Forest zone ( Table 5 and S4 Table).

Percentage out of range of the established haematological RIs compared to accompanying RIs
The percentage out of range (OOR) values of the study RIs were computed. Tables 6 and 7 show the proportion of normal Ghanaian adults whose haematology and biochemistry laboratory results would have been described as abnormal when the accompanying RIs provided by HORIBA (haematological) and DIALAB (biochemistry) are used.

Discussion
Although evidence suggests that disparities in socio-demographic and genetic factors affect the development of RIs, the healthcare system in Ghana continues to depend on pre-established RIs that were developed from Caucasian populations. This study reports the haemato-biochemical RIs from four eco-geographical zones in Ghana, highlighting gender-based differences and disparities in comparison to the RIs that accompany the analyzers used, as well as RIs from other studies. This study found the RIs for RBC, Hb, HCT, MCV and MCH to be higher in males compared to females whereas WBC and platelet counts were higher in females compared to males as consistent with previous reports in other African countries [8,[18][19][20][21][22]. The differences in RIs based on sex may be linked to the effect of menstruation and its associated increased demand for iron, differences in androgen hormones (testosterone and oestrogen) and the extent to which erythropoiesis and megakaryopoiesis are regulated in males and females [21,23,24].
Several variations between the RIs in this study and those provided by the manufacturer of the haematology analyzer were observed. The RIs for Hb level, HCT, MCV, MCH, MCHC were below the lower limit of the accompanying RIs. These parameters also presented with substantial misclassification, with %OOR ranging from 16.10%-35.8%, when compared to the accompanying RIs from the manufacturer. Other studies in Ghana [13] and other African countries [12,25,26] have highlighted similar disparities. This finding suggests that a proportion of normal participants whose haematological results are interpreted based on the preestablished (manufacturer's) RIs may be erroneously classified as having anaemia. The dropping of the lower limit of the red cell indices among the study population compared to the accompanying RIs may be attributed to the relatively lower ferritin and transferrin saturation among blacks [27] as well as poor nutritional status among the general Ghanaian population [28]. This highlights the influence of geographical variations on the normal physiology and support the need for the development of population-specific RI as mandated by the CLSI laboratory regulatory guidelines.
Furthermore, there were significant intra-population differences in the haematology RIs with respect to eco-geographic characteristics. The Transitional zone had the highest RBC, Hb and MCHC. This can be explained by the fact that the Transitional zone is mountainous (over 500 m above sea level) with conducive climate (optimal annual rainfall, temperature and humidity) and vegetation for agriculture. High altitude has been linked with increased erythropoiesis [29] and the predominant consumption of natural organic products (fruits and vegetables) compared to the other zones could also account for the relatively increased Hb. Furthermore, as consistent with our previous report [13], the Savannah zone had higher MON#, BAS# and BAS%. This finding could be linked to higher incidence of allergic conditions and parasitic infections in the Savannah zone [30].
This study also found significant gender-based and intra-population (based on eco-geography) variations in most of the biochemistry RIs. As coherent with previous studies in Ghana by Dosoo et al. [8] and Koram et al. [9] as well as studies by Abebe et al [31] in Ethiopia and Samaneka et al. [18] in Zimbabwe, the liver function analytes, kidney function analytes, and lipid profile (Trig and LDL-C) were higher among males compared to females. On the contrary, females had higher HDL-C compared to males. These findings were expected because males for instance have relatively greater skeletal muscle and bone mass which influences their serum creatinine. Additionally, the higher urea levels in males compared to females could be linked with the increased protein intake generally observed among males in the region. Of note, the Coastal Savannah had higher lipid profile whereas the Transitional zone had higher liver and kidney function analytes. These variations in biochemistry intervals could be attributed to differences in ethnic/genetic, nutritional and cultural/social lifestyle in the different

Limitations
This study is limited by the fact that lipid samples were non-fasting. Of note, according to most national and international guidelines, RIs for lipid profiles are established on fasting blood samples when used for cardiovascular risk assessment. Additionally, because the study was limited to young adult blood donors, this study was unable to determine the RIs of children and older individuals. Due to the difficulty in standardizing dietary patterns based on international guidelines, the study was unable to directly assess them. Further studies are thus warranted. Furthermore, the study was conducted among apparently healthy blood donors and may not be reflective of the general population. Interpretation of the results should thus be approached with caution.

Conclusion
This study established haematological and biochemistry RIs that would be potentially useful in the diagnosis, management and monitoring of disease progression in the study settings. There were variations in RIs within the population and between the study regions and other countries. Further local and nationwide studies are recommended to establish local and national RIs for haematological and biochemistry parameters.
Supporting information S1 File. Data (haematology) and Data (biochemistry): Excel sheet of datasets on which the conclusions of this manuscript were made. (XLSX) S1