**4. Should establishing separate normal ranges for African adolescents and pregnant women be considered?**

A number of studies similar to our published data [9], have reported age-related variation between male adolescents as compared to adults for Hb, Hct and RBC levels [25, 34, 45]. This observation is physiologically grounded on hormonal influence and as per the CLSI guidelines, partitioning reference intervals by age (or other subgroup considerations) may be appropriate. While these observations may not be of any medical significance, it should be taken into consideration whenever clinical trials target this population. To satisfy the statistical requirement for partitioning, there is need for further research on reference values among adolescents, as their participation in clinical trials increases.

Other than the RBC components mentioned above, no significant age differences have been observed in other laboratory parameters measured among males or in any parameters measured among females except for creatinine and BUN. Thus, for such parameters for which no differences are reported, adult values can be used in clinical trials involving adolescents.

With the advent of antiretroviral therapy for HIV and other interventions to improve maternal and child health, pregnant women and infants have become the focus of many health programs. However, few data exist regarding these important populations, despite increased clinical trials aimed at reducing mother-to-child HIV transmission. Although pregnancy-induced changes occur in hematological values including Hb, Hct and RBC count, very few laboratories provide specific reference ranges for pregnant women [63, 64].

In pregnancy, blood volume increases resulting in hemodilution. While the red cell mass increases during pregnancy, the plasma volume increases more resulting in a relative anemia. This leads to a lower Hb level, Hct and RBC. Hb is known to vary with gestational age with the highest values within the first and last trimesters and lowest during the second trimester. Similarly, the Hct and RBC decreases with gestational age. A stable higher upper reference limit for WBC count during pregnancy has been reported [65, 66]. WBC count is known to peak at delivery, thus limiting the use of this parameter as a marker for infection during delivery. This increase in WBC count results primarily from an increase in neutrophil counts and a slight increase in lymphocyte counts. Currently, there exists no African study designed to establish reference intervals during pregnancy and most laboratory information systems report reference values based on samples obtained from non–pregnant women which may not be useful for clinical decisions during pregnancy. Thus, there is an increased risk of overlooking important physiologic alterations resulting from pathological conditions and of misinterpreting normal changes as pathological events [64]. It is therefore important to develop reference intervals for women during pregnancy and the postpartum period for use in patient monitoring and management.

Laboratory Reference Intervals in Africa 315

**Division of AIDS (DAIDS) toxicity grading** 

**Comparison Grade 1 Grade 2 Grade 3 Grade 4** 

**percentile n % n % n % n % n %** 

**MGH USA reference** 

**Hemoglobin Males (**g/dl**)** 140 13.5-17.5 65 46 2 1.3 0 0 2 1.3 0 0

**(**g/dl**)** 153 12-16 61 40 11 7.9 8 5.7 14 10 0 0

**Platelets (**109 cells/L**)** 293 150-350 53 18 6 2 6 2 0 0 0 0 **WBC** (109 cells/L) 293 4.5-11.0 66 23 2 0.7 0 0 0 0 0 0

(109 cells/L) 293 1.0-4.8 6 2 0 0 0 0 0 0 0 0

(109 cells/L) 293 1.8-7.7 110 38 25 8.5 9 3.1 1 0.3 0 0 **Eosinophil** (109 cells/L) 293 0-0.5 130 44 60 20.5 12 4.1 0 0 0 0

**ALT (SGPT)** (U/ L) 293 0-35 30 10 12 4.1 1 0.3 0 0 0 0 **AST (SGOT)** (U/ L) 293 0-35 40 13 9 3.1 3 1 0 0 0 0 **Total Bilirubin (**μmol/L**)** 293 5.1-17.0 90 30 37 12.7 27 9.2 4 1.4 1 0.3 **Creatinine (**μmol/L**)** 293 0-133 4 1 4 1.4 0 0 0 0 0 0

**\*CD4 (**Cells/ μl**)** 293 404-1612 6 2 3 1 1 0.3 0 0 0 0

\*Reference ranges provided by Becton-Dickinson with the MultiTEST IMK Kit Reagent package (12/2000;23-3602-02) -

**Table 4.** Frequency of adverse events and out of range values comparing western Kenyan cohort to

While it is desirable to generate reference intervals for different populations, the procedure remains a challenge due to the prohibitive cost involved in performing these studies and the limitation in identifying suitable healthy reference individuals. Thus, the CLSI recommendation that all diagnostic laboratories should determine and maintain their own reference interval for each laboratory parameter is impractical. The revised CLSI guidelines

DAIDS- Division of AIDS tables for grading the severity of adult and pediatric adverse events [26] - MGH-

**(25th percentile)[6]** 

**intervals** 

**out of range** 

**Hct (females) (**%**)** 140 36-46 74 53 **Hct (males) (**%**)** 153 41-53 88 58 **RBC (males) (**1012 cells/L**)** 140 4.5-5.9 29 19 **RBC (females) (**1012 cells/L**)** 153 4.0-5.2 32 23 **MCV (**fL**)** 293 80-100 157 54

**Basophil count** (109 cells/L) 293 0-0.2 5 2

(109 cells/L) 293 0-0.8 0 0

**Glucose** mmol/L 293 4.2-6.4 210 71 **BUN (**mmol/L**)** 293 3.6-7.1 246 84

**\*CD8** (Cells/ μl) 293 220-1129 13 4

Massachusetts General Hospital weekly case records [25]

DAIDS and North American derived MGH values

**Parameter n 95th**

**Hemoglobin Females** 

**Lymphocyte count** 

**Neutrophil count** 

**Monocyte count** 

**6. Conclusion** 

## **5. A case for African/ Region specific toxicity tables?**

Under a research-based approach, applying the US Massachusetts General Hospital derived reference intervals to our reference population from western Kenya during screening for a clinical trial (Table 4), over 58% of the volunteers would have been excluded from the trial despite having laboratory results consistent with the general population from which they were derived. This erroneous screening out of otherwise healthy volunteers would have important implications on study costs, work load and time, as more volunteers would be need to be screened in order to meet the required target [15].

Similarly, applying the DAIDS toxicity tables to our population, some of our calculated reference intervals fall between the normal, and grade 1–2 toxicity grading in the DAIDS system (Table 4). Using the clinic based approach, 40% of our otherwise healthy study participants would have erroneously been considered to have at least one laboratory-based grade 1–4 toxicity adverse event. The lower range for Hb, neutrophil counts, as well as the upper range for eosinophil counts and bilirubin would be considered as grade 2 adverse events, for example. Even though studies have documented these findings, this information is not widely known and as a result, DAIDS has issued only 1 set of "standard" toxicity tables without considering racial or ethnic differences [57]. Thus, during international clinical trials, these tables are used as guidelines in the conduct of such trials. This may result in a situation where the results of a clinical trial cannot be generalized to the population in question since a majority of otherwise healthy participants are screened out. Moreover, given that the investigational product is intended for use within the same population being sampled, this may complicate post-market analysis or application of the product for the general population. Unfortunately, there are no comparable tables from Africa on which such clinical decisions can be based. It is therefore important that African countries carry out large studies in different regions of Africa for such parameters to establish African toxicity tables.


\*Reference ranges provided by Becton-Dickinson with the MultiTEST IMK Kit Reagent package (12/2000;23-3602-02) - DAIDS- Division of AIDS tables for grading the severity of adult and pediatric adverse events [26] - MGH-Massachusetts General Hospital weekly case records [25]

**Table 4.** Frequency of adverse events and out of range values comparing western Kenyan cohort to DAIDS and North American derived MGH values
