Ethical approval was obtained from the Cameroon National Ethical Committee of Research for Human Health (permit number: 2016/01/701/CE/CNERSH/SP). Every participant gave written informed consent before the sample collection.

This was a cross-sectional study conducted from November 2015 to September 2016. The target population constituted blood donors, because they were submitted to a rigorous medical screening that took into account several criteria for detecting pathology, thus ensuring a healthy population for the study. The recruitment sites were the blood banks of the most frequented hospitals, including the University Teaching Hospital of Yaoundé, the Yaoundé Central Hospital, and the Hospital Centre of Essos. A stratified and clustered sampling strategy enabled us to divide the population into two groups (men and women) and collect around 50 samples per site and per sex. According to the Clinical and Laboratory Standards Institute,⁴ a study requires a minimum of 120 samples in each group in order to obtain valid reference values. A survey was administered before the blood donation that enabled the selection of participants and the collection of socio-demographic data. All of the eligible blood donors who gave their informed consent were included in the investigation. People who met one or more of the following criteria were excluded: (1) currently sick or being followed up for a known pathology, (2) pregnant women, (3) undergoing treatment for any disease, (4) history of blood donation or transfusion within the immediate past 4 months, (5) regular smokers and (6) hospitalisation within the immediate past month. In addition, we excluded people with a positive serology result for any one of the serological tests carried out at the blood bank, including HIV, syphilis, hepatitis B virus, hepatitis C virus, positivity to blood parasites through a thick or thin blood film, platelet aggregates, C-reactive protein > 6 mg/L, or inflammatory syndrome, as well as people with hyperthermia (˃ 37.8 °C) or hypothermia (< 35 °C) as determined by thermometer as a result of temperature control-related conditions.

For each participant, blood was drawn from a superficial vein in the upper limb using venepuncture sampling. This allowed the collection of 5 mL of blood in an ethylene diamine tetraacetic acid tube, which was used for full blood count, reticulocyte count and blood parasite analysis by a thick and thin blood film, and 4 mL of blood in a tube without anticoagulant for C-reactive protein analysis. Those tests were performed at the laboratory of Centre Pasteur Cameroon. Samples were stored at room temperature in an air-conditioned room (20 °C – 23 °C). Samples were transported in a cooler containing a tube rack to avoid thermal or mechanical shocks that could damage cells and cause haemolysis. Analyses were done within 6 hours of blood collection. A peripheral blood smear and a thick film were prepared for every sample.

Serological tests for HIV, syphilis, hepatitis B virus, and hepatitis C virus were carried out by every blood bank. Human immunodeficiency virus (HIV) antibodies were detected by two rapid tests that detect both HIV-1 and HIV-2 infections: (1) the ImmunoComb HIV 1&2 BiSpot (Orgenics, Courbevoie, France) and (2) the Determine HIV-1/2 (Abbott Laboratories, Chicago, Illinois, United States), an immunochromatographic assay. Samples reactive to only one of the two rapid testing procedures were considered indeterminate and further tested using an ELISA method. Hepatitis B virus was detected using a one-step, immunoassay-based DIASpot HBsAg test kit (DIASpot Diagnostics, Jawa Barat, Indonesia) for qualitative detection of serum hepatitis B surface antigen. IgG antibodies to hepatitis C virus were detected using DIASpot HCV-Ab test strips (DIASpot Diagnostics, Jawa Barat, Indonesia), an immunochromatographic assay. Syphilis was diagnosed using the venereal disease research laboratory test and the Treponema pallidum hemagglutination assay test (Omega Diagnostic, Alva, Scotland, United Kingdom).

The full blood and reticulocyte counts were analysed using Pentra DX Nexus (Horiba-ABX, Montpellier, France). This analyser allows for the differentiation and quantification of 28 parameters of a haematopoietic population with five reference methods of measurement: (1) flow cytometry, (2) cytochemistry, (3) impedance, (4) absorbance and (5) optical cytometry. The reticulocyte count was performed using fluoro-flow cytometry with orange thiazol.

Levels of internal quality controls (Horiba-ABX difftrol, ABX minotrol reti) were analysed daily on three samples: (1) low, (2) normal and (3) high. The performance of the laboratory was checked based on the external proficiency testing program with Biology Prospective (Villers-lès-Nancy, France). The laboratory has been audited as per the World Health Organization Regional Office for Africa’s Laboratory Improvement Process Towards Accreditation (SLIPTA) Checklist, and has achieved a four-star ranking on the SLIPTA Tier of Recognition of Laboratory Quality Management (valid: September 2017 to September 2019).

Data were registered in the Excel 2013 software (Microsoft Corp., Redmond, Washington, United States) and analysed with Epi Info 7 (Centers for Disease Control and Prevention, Atlanta, Georgia, United States) and SPSS 19 software (IBM Corp., Chicago, Illinois, United States). The data sets were not tested for a Gaussian distribution, which was taken into account in the data analysis. For each parameter, the standard deviation and the median with the 2.5th percentile and 97.5th percentile were calculated. The 2.5th percentile was used to define the lower reference limit, while the 97.5th percentile defined the upper one. A Kruskal–Wallis test for the two groups was used to correlate parameters according to sex. The threshold for significance was set at p < 0.05 for all statistical analyses.

The total number of people approached and asked to participate in the study was 375. After applying the exclusion criteria, 294 participants (161 men and 133 women) were maintained in order to determine normal values for the local population. A positive serology result for any one of the serological tests carried out at the blood bank, and positivity to blood parasites were major reasons participants were excluded (Table 1). The sex ratio was 1:12. Their ages ranged from 18 to 55 years for men and 18 to 53 years for women (Figure 1). Most participants identified as southern forest peoples (Table 2). The majority of the population (59%) was between age 18 and 28 years and 31% was between age 29 and 39 years. In terms of blood donation, 61% of the population had never donated blood before. The distribution was approximately equal between the three sites. Once the exclusions had been applied, around 45 donors of each sex were included per site.

The median haemoglobin concentration was 135 g/L in men and 114 g/L in women and the difference was statistically significant (p < 0.001) (Table 3). A statistically significant difference between men and women was also noted for red blood cell and haematocrit values, with men having higher values than women. Normal local values were found for erythrocyte parameters; no differences between men and women were observed for mean corpuscular volume, the mean corpuscular haemoglobin concentration, the mean haemoglobin concentration, and the red blood cell distribution index. The reticulocyte count was higher in men compared with women and the difference was statistically significant.

The number of platelets was higher in women (median: 243 × 10⁹/L) than in men (median: 211 × 10⁹/L) (Table 4). Neither the mean platelet volume nor the platelet distribution index differed by sex.

The median white blood cell count was 4.1 × 10⁹/L in men and 4.6 × 10⁹/L in women (Table 5). Statistically significant differences between men and women were found in the levels of neutrophils, eosinophils and monocytes. No differences between men and women were observed for lymphocytes or basophils.

Haematological values of the local population are necessary in order to improve daily diagnosis. We conducted a small-scale pilot study due to limited financial support before launching a project. Although the size of our sample was smaller than in other studies, the median values of many parameters, especially haemoglobin and platelet counts, were comparable to the ones found in other studies conducted in Africa but different to those in the northern hemisphere countries. This study took place in the capital city, which is a cosmopolitan town. It would be more appropriate to use these results rather than the standards from foreign countries and to eliminate the variability in reference ranges between local laboratories.