Factors specified when reference values are established include: (1) Make up of reference population in terms of age, gender and genetic and socio-economic factors. (2) The inclusion and exclusion criteria used. (3) The conditions, both physical and physiological under which the reference population is sampled and studied. (4) The procedure of collecting the specimen, including how the subject was prepared before collection and (5) the method of analyzing the sample used giving details of its precision and accuracy [2]. Since measured biochemical parameters are affected not only by individuals’ factors such as age, sex, diet but also by population and ecological factors such as ethnic background, climate, geography and altitude, they are found to vary not just between individuals but between populations as well [3]. Presently most clinical laboratories in Meru County as well as the whole republic of Kenya use reference values as indicated on reagent kits or those that are published in medical and laboratory textbooks to interpret patient results. This can be of a grave mistake bearing in mind that parameters vary from region to region [4]. The patient’s results can also vary due to the methods used in analyzing, that is either manual or by automation. Thus, it is important that clinical chemistry laboratories determine reference values that are specific to them are accurate and precise, there is need to establish reference values following the standard populations they serve. To ensure the results operating procedures (SOP) [5].

This study is designed to establish health associated reference ranges to be used by clinical chemistry laboratories in the area of study and determine whether there are significant variations between them and those used in the facilities as provided by the reagent manufacturers. 

MATERIALS AND METHODS

Selection of reference population

The reference population was selected based on the guidelines described by the CLSI, 2000. According to the guidelines, the reference individuals selected should be closely similar to the patient population under study and should be of the same age to be clinically significant. A priori sampling method was employed in this study. This is where selection of an adequate number of subjects who serve as reference individuals takes place first and then samples are drawn for analysis. Children between 1-17 years of age were randomly selected after holding community meetings with the leaders and parents/guardians to explain the objectives of the study.

Study design

This was a population based cross-sectional study involving 740 healthy male and female subjects of age 1 to 17 years.

Specimen collection

5 ml of blood was collected by venipuncture using a 23 gauge butterfly needle with a 5 ml syringe after sterilizing the area with 70% alcohol. The blood was dispensed into a plain vacutainer (without anticoagulant) tube then transferred in 2 ml tubes (vacutainer TM, Becton Dickinson, Franklin Lakes, NJ). Each of the tubes containing the specimens was labeled clearly with the subject’s name, the study number and the date of collection of the sample. The specimens were arranged in Styrofoam cool boxes at 4°C and covered to protect them from heat and sunlight, awaiting transportation to the main analytical center

Specimen transportation, processing and storage

The specimens collected in the field were ferried from the point of collection to the clinical chemistry laboratory of Meru Level Five Hospital in cool boxes within 2 h of collection at room temperature. Upon arrival at the clinical chemistry laboratory, the blood specimens were centrifuged at a speed of 3000 rpm for two minutes to obtain serum. The serum was transferred to separate tubes labeled with subject’s identification details. Laboratory analysis was done as soon as possible to avoid loss of sample viability. If analysis was not done immediately, the samples were stored at -20°C for a period not exceeding seven days.

Laboratory analysis

Serum samples were initially screened for human immunodeficiency virus types 1 and 2 (HIV-1 and HIV-2) by using HIV 1/2 Stat-Pak^®, Chembio Diagnostic Systems that detects anti-HIV antibodies in human serum. HBsAg one step hepatitis B Surface Antigen Test Strip (HBsAg, Beijing, China) was used for screening hepatitis by a qualitative lateral flow immunoassay test. Screening for syphilis antibodies was by a syphilis ultra-rapid test using one step anti-Treponema pallidum test strip (Anti-TP Test Strip, China). Five liver biochemistry tests were carried out on the serum samples: sodium (NA), potassium (K), chloride (CL), urea (UREA) and creatinine (CRE). All the assays were carried out as per the standard operating procedures (SOPs) followed at the hospital laboratory using DRI-CHEM NX500i dry chemistry clinical analyzer (Fujifilm, Europe).

Reagent preparation

The machine uses coded slides that are commercially acquired and specifically tailored for the equipment (Fuji DriChem slides). Each slide is impregnated with the reagent for a specific parameter and is labeled clearly. The slides for the various tests being carried out are inserted in the machine along with the sample and once the patient details are entered and the machine set to start, the tests run automatically and a print-out of the results is obtained.

Calibration of the test

Calibration was done using a magnetic card called QC card that comes with every reagent box. The card is passed through the QC card reader in the analyzer whenever slides from a new lot number are being used.

Quality control (QC) materials

To ensure accuracy and precision of the test results, all pre-analytical, analytical and post analytical precautions were taken into consideration. Internal QC materials from Roche diagnostics; Precinorm and Precipath, were run daily or at any other time as deemed necessary. External QC materials were from BioRad, (Lyphochek^® Unasssayed Chemistry Quality Control (USA). The multisera were supplied in lyophilized form and were reconstituted when being used as per the manufacturer’s guidelines.

DATA MANAGEMENT AND STATISTICAL ANALYSIS

Data was categorized by sex and each parameter was then examined as a histogram and as a normal probability plot. Kolmogorov-Smirnov Test was performed with significance at p=0.05 level to test fit of the data to Gaussian distribution. Since the results of the Kolmogorov-Smirnov test were not significant (p>0.05), the data was found to assume a normal distribution and was further subjected to parametric statistical methods. The lower and upper reference interval limits for each parameter were calculated from the arithmetic mean (X) ± 1.96 times standard deviation (X ± 1.96 SD) to obtain the 2.5 and 97.5% percentiles. Students T-test was used for comparison of means between sexes. The tests were conducted at 95% confidence interval and significance level of 5%. P-values of less than or equal to 0.05 (p ≤ 0.05) were considered statistically significant. The performance of analytical instruments and methods to analyze the levels of the selected analytes were achieved by using the paired T-test. Quality control was observed throughout the study to make sure that all the results were in the recommended ranges of reporting. This was achieved by use of normal and pathological (PNU and PPU) pre-determined values, respectively.

RESULTS

Sex-specific reference values for CRE, UREA, NA, K and CL 

From the study, it was noted that reference values that are used in management of children differ with those that were established, therefore indicating that there is need to have each laboratory determine reference values that are specific to its population as required by the CLSI. 740 samples were analyzed in this study out of the 768 initially collected for the study. 28 samples were excluded because 6 (21.4%) of them were HIV positive and 22 (78.6%) were hemolyzed. Reference values for the five biochemical parameters were established for both males and females of age 1 to 17 years. Construction of these reference values was by use of the 2.5^th and 97.5^th percentiles as lower and upper limits respectively and at 95% confidence interval. This is in accordance with CLSI guidelines for reference interval determination [3].

Table 1 shows sex-specific reference values for every parameter depending on the p-values obtained from the difference between male and female subjects. Significant sex differences were observed in NA (p=0.003) and K (p=0.009). Male adolescents had higher values for UREA and NA than female adolescents. K demonstrated higher values compared to manufacturers’ ranges. This could be due to differences in diet, genetics and analytical methods. Sex differences observed for NA and K may be due to differences in response to dietary salts brought about by the effects of sex hormone patterns and sex-related genetic factors. CL values remained relatively constant between male and female subjects but exhibited higher upper reference limit values than those reported in literature. This could be attributed to diet and geographical location.

In the course of the study, control value results as well as the standard deviation (SD) from the value were recorded every day. Table 2 compares reference range values established with those found in the literature. This was done by comparing the values for the lower and upper reference limits as well as the interval values for each analyte.

From the study, it was noted that significant differences exist between the reference ranges developed and those in use at the hospital. Out of the five parameters studied, only creatinine was found to have distinct values for adults and children whereas for the other parameters, the same reference values were being used when interpreting test results for both children and adults.

DISCUSSION

The results of this study provide the pioneer clinical chemistry reference ranges for the population of 1 to 17 years in Meru County, Kenya using 740 samples, 380 males and 360 females. The number of participants in each category was more than the minimum number of 120 participants per subgroup required to determine reference ranges as recommended by CLSI 2000. The tests were done using the same analytical methods and results expressed in the same units as those found in literature for easy comparisons. External and internal quality control methods were closely followed and monitored throughout the study [6] in addition to following the SOP at Meru Level Five Hospital.

In the study, it was established that there exists no reference ranges in Meru County for children and adolescents of 1-17 years for four out of the five parameters studied. During clinical trials, this group of the population was found to be considered together with the adult population. However even for CRE, the reference values found in use are not age specific. There was no reference range values available for Urea, K, Na and Cl therefore the reference values developed in this study were compared against those for the adult population that was being used in hospital. Assuming that adults and children exhibit the same reference ranges is a big mistake since reference ranges are known to vary with age among many other factors. CRE had both lower and upper reference limits that were lower than the manufacturers’ limits.

Generally, reference ranges have been shown to vary between different populations due to differences in genetics, physical, environmental and socio-economic conditions and diet [7]. The reference values for the parameters analyzed in this study differ from those used to service the population. This clearly indicates the necessity to determine sex- and age-based reference values for specific populations instead of taking a set of reference values determined for one population and use them on another population. This will decrease the frequency of values reported as abnormal in otherwise healthy children and adolescents.

CONCLUSION

From the study it was observed that no reference ranges are available for the study population in Meru County, Kenya and physicians and other health workers rely on adult values available from some Caucasian populations to interpret laboratory results for this group of population. This is a grave mistake because serum biochemistry parameters are known to vary with age, thus adult values cannot adequately represent children reference ranges. The importance of serum biochemistry normal ranges in the diagnosis and monitoring of disease cannot be underestimated; therefore establishment and use of local reference ranges should be encouraged because it enhances patient care and health research [8].

The results of this study show that reference values obtained vary with those from literature and those that are used at Meru Level Five Hospital. This indicates the need to determine population specific ranges instead of using a general range developed using a different population. There was little or no information found in literature for this population thus comparisons for four out the five parameters studied were done with those of adult populations [9].

It is recommended that similar studies of children in Africa should be carried out so as to broaden the present findings thus enabling improved care and conduct of clinical trials. Population-specific reference ranges/values obtained will be useful to achieve accurate and clinically relevant results that will provide true information about the patient’s state of health in the region.

• 
  Table 1
• 
  Table 2