Correlation Between Total Lymphocyte Count, Hemoglobin, Hematocrit and CD4 Count in HIV Patients in Nigeria
Charles Iheanyichi Emuchay,
Shemaiah Olufemi Okeniyi
Joshua Olusegun Okeniyi
The expensive and technology limited setting of CD4 count
testing is a major setback to the initiation of HAART in a resource limited
country like Nigeria. Simple and inexpensive tools such as Hemoglobin (Hb) measurement
and Total Lymphocyte Count (TLC) are recommended as substitute marker. In order
to assess the correlations of these parameters with CD4 count, 100 apparently
healthy male volunteers tested HIV positive aged ≥20 years but ≤40
years were recruited and from whom Hb, Hct, TLC and CD4 count were obtained.
The correlation coefficients, R, the Nash-Sutcliffe Coefficient of Efficiency
(CoE) and the p-values of the ANOVA model of Hb, Hct and TLC with CD4 count
were assessed. The assessments show that there is no significant relationship
of any of these parameters with CD4 count and the correlation coefficients are
very weak. This study shows that Hb, Hct and TLC cannot be substitute for CD4
count as this might lead to certain individuals deprivation of required
Received: March 17, 2013;
Accepted: May 15, 2013;
Published: November 26, 2013
The number of people living with HIV globally continues to grow, a large number
of these lives in developing and resource limited countries, like Nigeria, with
millions of people in these countries requiring life sustaining highly active
anti retroviral therapy (HAART) (Alavi et al., 2009;
UNAIDS/WHO, 2005, 2006, 2007,
2009; WHO, 2004; Kumarasamy
et al., 2002). In the resource-limited countries, the high cost of
widespread and routine use of CD4+ T lymphocytes count and plasma viral load
testing in the management of HIV infection has militated against its usage (Daryani
et al., 2009; Solomon and Solomon, 2004;
Kumarasamy et al., 2002). The World Health Organization
(WHO) suggested that this limitation could be tackled using simple tools such
as hemoglobin (Hb) measurement and TLC as markers to initiate HAART and as a
substitute marker to monitor immune response to therapy in symptomatic HIV patients,
in the resource-limited settings (WHO, 2004). There exist
different arguments by studies based on this suggestion (Akinola
et al., 2004; Spacek et al., 2003).
The expensive and technology limited setting attending CD4 count testing is
major setback to the initiation of HAART. It is therefore resolved, in this
study, to look into the capability of TLC, Hb and hematocrit (Hct) as substitute
for CD4 count in HIV/AIDS patients through the correlation assessments of these
parameters as alternate for CD4 count to initiate HAART in HIV/AIDS patients
MATERIALS AND METHODS
Subject: One hundred apparently healthy male volunteers
tested HIV positive, aged ≥20 years but ≤40 years, were recruited from
randomly selected sites throughout South East Nigeria. These are yet to receive
HAART. Exclusion criteria include known diagnosed diseases, use of chronic medication,
oral temperatures>37°C, free from the use of interfering or toxic drugs.
Most importantly, women are also excluded from the study because of their complication
from menstrual cycle, pregnancy and lactation and other hormonal complications.
All these constitute methods of subjects selection described by Oosthuizen
et al. (2006).
Blood collection and HIV serology: Whole blood was collected with a
Vacutainer system in 10 mL tubes containing EDTA. HIV status was determined
with plasma samples by an enzyme immunoassay (EIA) with ImmunoComb®
(Orgenics Isreal) and Determine test (Abbott). Blood sample for other analysis
were obtained based on the instruction given in the manual of the analyzer used.
Hematological analysis: QBCTM AutoreadTM Plus
Centrifugal Hematology System was used for the whole blood analysis of hematological
parameters. It utilizes precision-bore glass tubes pre-coated with potassium
oxalate, acridine orange fluorochrome stain and an agglutinating agent. QBC
tubes made specifically for capillary blood (finger-stick samples) additionally
contain a coating of anti-coagulants. During high-speed centrifugation of the
blood-filled tube, the cells form in packed layers around the float, which has
descended into the buffy coat. The spun tube is inserted in the QBC Autoread
Plus Analyzer, where it is automatically scanned and fluorescence and absorbance
readings are made to identify the expanded layers of differentiated cells. Volumes
of these packed cell layers are then computed to obtain quantitative values.
The obtained quantitative values include Hematocrit (Hct) (in percent), Hemoglobin
(Hb) (in gram per deciliter), Platelet Count (Pl) (number of cellx109
L-1), White Blood Cell Count (leucocytes/WBC) (expressed in number
of cellx109 L-1). Other quantitative values obtained include
Granulocyte Count (Gr) (in percent and number of cellx109 L-1)
and Lymphocyte-Monocyte Count (TLC) (in percent and number of cellx109
CD4+ T Lymphocyte analysis: Quantification of human peripheral
blood CD4+ T lymphocytes were carried out using Dynal®
T4 Quant Kit. This utilizes immunomagnetic cell separation EDTA anti-coagulated
blood samples to enable a rapid and direct quantification of CD4+
T lymphocytes. The cell isolation takes place at room temperature by depletion
of monocytes from blood sample using Dynabeads CD14, isolation of CD4+
T cells from monocyte-depleted blood using Dynabead CD4, washing of bead bound
cells to remove contaminating cells and counting of CD4+ T cells.
Statistical analyses: Initial distribution of descriptive statistics
for each hematological characteristic in the reference subjects was analysed
using STATGRAPHICS® Centurion XVI version 16.1.11 by StatPoint
Technologies, Inc. Also, linear correlation models of each of the hematological
parameters and those of the combined parameters with CD4 count were analyzed
and investigated using personally developed Visual Basic® program
version 6.0. These were run on hp pavilion dv6-2150us (2009 model) with Intel®
CORE i3 CPU M330 @ 2.13 GHz (Okeniyi and Okeniyi,
2012). Fittings of correlation models were studied using the correlation
coefficient, R and the Nash-Sutcliffe Coefficient of Efficiency (CoE) given
respectively as (Omotosho et al., 2011; Krause
et al., 2005):
where, Oi and Pi are respective observed and predicted
CD4 count, while
are the respective mean of the observed and predicted CD4 count.
Ethics: Informed consent was obtained from each participant.
RESULTS AND DISCUSSION
The mean, Mean±2SD, median and 95th-percentile confidence interval for
the hematological parameters and CD4+ T cells of the recruited 100
subjects (between 20≤age (years) ≤40) that were tested HIV positive and
are yet to receive HAART, are shown in Table 1. From the table,
it could be observed that the mean values of hematological parameters obtained
in this study bear similarity with those from other African studies (Akinola
et al., 2004; Akanmu et al., 2001).
Hct value of 33.6±5.97% from 100 male patients in this study compares
well with 34.1±7% obtained from 64 male patients in Akinola
et al., 2004), even as the WBC of 4.2±1.38 (x109
L-1) in this study falls within the range of 4.788±2090 (x109
L-1) obtained in that study. Also, CD4 count of 228±140.66
μL-1 from this study bear good comparisons with 222±189
μL-1 obtained in Akinola et al.
(2004) and the 233 μL-1 obtained by Akanmu
et al. (2001). However, CD4+ T cells value of 279±225
μL-1 obtained from patients living in Iran, Asia (Alavi
et al., 2009), is significantly higher than the results from this
and other studies of patients living in Nigerian, Africa (Akinola
et al., 2004), even though hemoglobin and hematocrit exhibited lower
mean values in the former than in the latter studies. However, the lymphocyte
count of 3.3±1.1(x109 L-1) obtained in this study
is high compared to lymphocyte counts obtained in the other studies mentioned
|| Measured parameter for HIV positive male adults in South
|NS: Not significant
|| Linear correlation of CD4 cells with other hematological
|aCases<cutoff with CD4<200
The correlation coefficient, R and the p-value of the slope, which is also
that of its ANOVA, obtained from linear relationships of each of the hematological
parameters with CD4 count are shown in Table 2. Also shown
in the table are cases found in each parameters cutoff and those greater
than the cutoff of CD4 count. From this, the p-values obtained from the ANOVA
for each of the parametric correlations are greater than 0.05 showing that there
is no statistically significant relationship between any of the parameters and
CD4 at the 95.0% or higher confidence level (Table 2). Although
CD4 count exhibited positive correlation with hemoglobin (R = 15.38%) and hematocrit
(R = 12.71%) but negative correlation with lymphocyte count (R = -10.57%), these
R-values indicate relatively weak relationship between these variables and CD4
count. This agrees with Akinola et al. (2004)
which is also a study based on patients living in Nigerian. However, results
from this study bear disagreement with that obtained in (Alavi
et al., 2009), save for the lack of fit encountered with the hemoglobin
and the hematocrit, as well as that from (Spacek et
al., 2003) both of which were not based on patients from Africa. Suggestions
from Lugada et al. (2004) identified causes
for these kind of disagreements as that which could be due to environmental
and genetic factors.
The combination of all these hematological parameters exhibit linear correlation
with the CD4 cells which could be represented as:
For this, the correlation coefficient R = 0.1848, Nash-Sutcliffe CoE = 0.0342
and p-value = 0.3402. All these estimates of model validation parameters further
reinforced that the correlation between CD4 counts and the studied hematological
parameters are very poor. Specifically, the correlation coefficient showed that
only 3.42% of the variability in the CD4 count could be accounted for by the
fitted correlation model, while the CoE statistics showed that the observed
model of CD4 count would be better than the predicted model from the fit. Also,
the p-value model showed that there is not a statistically significant relationship
between the hematological variables and CD4 count at the 95.0% confidence level.
This bear dire implications, especially, from the view of TLC, which has been
recommended to stand as substitute marker for CD4 count for resource limited
countries to initiate HAART (WHO, 2004). Out of the 100
HIV positive individuals, two (2) fell in the ≤1.2x109 L-1
cutoff of the TLC recommendation while 40 that had CD4 <200 cells μL-1
had TLC >1.2x109 L-1. It has been reported that with
results like these, certain number of HIV positive individuals would have been
deprived of needed treatment (Akinola et al., 2004)
for their TLC result might have been used for HAART initiation going by WHO
recommendation. Results from the patients studied in this work show that neither
Hemoglobin nor Hematocrit, not even TLC, can be substitute for CD4 count of
the patients. Thus, further research work are required for patients in Nigeria
and in other resource limited countries in Africa, investigating parameters
which would bear correlation with CD4 count for adequate indication of costly
CD4 in resource limited countries.
Correlation between CD4 count and hematological parameters such as total lymphocyte
count, hemoglobin and hematocrit had been studied in this work. From this, conclusions
that could be drawn include:
||The mean values of hematological parameters obtained in this
study bear similarities with those obtained in other studies for patients
living in other part of Nigeria (Africa)
||Individual correlations of total lymphocyte count, hemoglobin and hematocrit
with CD4 showed that all of the parameters exhibit poor correlation with
CD4 count, which were attendant with low correlation coefficients and non
significant ANOVA p-values
||Combined correlation of total lymphocyte count, hemoglobin and hematocrit
with CD4 showed that the resulting correlation fitting function exhibited
poor correlation coefficient, R = 0.1848 and poor Nash-Sutcliffe Coefficient
of Efficiency, CoE = 0.0342 and non-significant p-value = 0.3402
||These results lead to the conclusion that hematological parameters such
as total lymphocyte count, hemoglobin and hematocrit are not suitable for
representing CD4 even as suggestion for further study are recommended for
finding suitable alternative for indicating costly CD4 count in limited
1: Akanmu, A.S., I. Akinsete, A.O. Eshofonie, A.O. Davies and C.C. Okanny, 2001. Absolute lymphocyte count as surrogate for CD4+ cell count in monitoring response to antiretroviral therapy. Niger. Postgraduate Med. J., 8: 105-111.
PubMed | Direct Link |
2: Akinola, N.O., O. Olasode, I.A. Adediran, O. Onayemi and A. Murainah et al., 2004. The search for a predictor of CD4 cell count continues: Total lymphocyte count is not a substitute for CD4 cell count in the management of HIV-infected individuals in a resource-limited setting. Clin. Infect. Dis., 39: 579-581.
CrossRef | Direct Link |
3: Alavi, S.M., F. Ahmadi and M. Farhadi, 2009. Correlation between total lymphocyte count, hemoglobin, hematocrit and CD4 count in HIV/AIDS patients. Acta Medica Iranica, 47: 1-4.
Direct Link |
4: Daryani, A., M. Sharif, M. Meigouni, F. Baba Mahmoudi and A. Rafiei et al., 2009. Prevalence of intestinal parasites and profile of CD4+ counts in HIV+/AIDS people in North of Iran, 2007-2008. Pak. J. Biol. Sci., 12: 1277-1281.
CrossRef | Direct Link |
5: Krause, P., D.P. Boyle and F. Base, 2005. Comparison of different efficiency criteria for hydrological model assessment. Adv. Geosci., 5: 89-97.
Direct Link |
6: Kumarasamy, N., T.P. Flanigan, A.P. Mahajan, C.C. Carpenter, K.H. Mayer and S. Solomon, 2002. Monitoring HIV treatment in the Developing world. Lancet Infect. Dis., 2: 656-657.
PubMed | Direct Link |
7: Lugada, E.S., J. Mermin, F. Kaharuza, E. Ulvestad and W. Were et al., 2004. Population-based hematologic and immunologic reference values for a healthy ugandan population. Clin. Diagn. Lab. Immunol., 11: 29-34.
8: Okeniyi, J.O. and E.T. Okeniyi, 2012. Implementation of kolmogorov-smirnov p-value computation in visual basic®: Implication for Microsoft Excel® library function. J. Stat. Comput. Simul., 82: 1727-1741.
CrossRef | Direct Link |
9: Omotosho, O.A., C.A. Loto, O.O. Ajayi and J.O. Okeniyi, 2011. Statistical-based performance evaluation of inhibitors on the corrosion of steel rebar in concrete. Botswana J. Technol., 19: 220-226.
10: Oosthuizen, W., A. van Graan, A. Kruger and H.H. Vorste, 2006. Polyunsaturated fatty acid intake is adversely related to liver function in HIV-infected subjects: The THUSA study13. Am. J. Clin. Nutr., 83: 1193-1198.
Direct Link |
11: Solomon, S.S. and S. Solomon, 2004. A comparison of the delivery of antiretroviral therapy system between Brazil and India. JK Sci., 6: 168-171.
Direct Link |
12: Spacek, L.A., M. Griswold, T.C. Quinn and R.D. Moore, 2003. Total lymphocyte count and hemoglobin combined in an algorithm to initiate the use of highly active antiretroviral therapy in resource-limited settings. AIDS, 17: 1311-1317.
Direct Link |
13: UNAIDS/WHO, 2005. AIDS epidemic update. Joint United Nations Programme on HIV/AIDS, Geneva, Switzerland.
14: UNAIDS/WHO, 2006. AIDS Epidemic Update. Joint United Nations Programme on HIV/AIDS Geneva. http://data.unaids.org/pub/epireport/2006/2006_epiupdate_en.pdf.
15: UNAIDS/WHO, 2007. AIDS epidemic update. Joint United Nations Programme on HIV/AIDS Geneva. http://data.unaids.org/pub/epislides/2007/2007_epiupdate_en.pdf.
16: UNAIDS/WHO, 2009. AIDS epidemic update. Joint United Nations Programme on HIV/AIDS Geneva. http://data.unaids.org/pub/report/2009/jc1700_epi_update_2009_en.pdf.
17: WHO, 2004. Scaling up antiretroviral therapy in resource-limited settings: Treatment guidelines for a public health approach. World Health Organization, Geneva, Switzerland, pp: 1-68.