The typical pattern of HIV primary infection is characterized by high levels of virus in blood followed by a progressive loss of CD4 T cells, elevation of CD8 T cells and progressive impairment of T cell functions. Between the primary infection and the final progression to AIDS there is a period of clinical latency lasting 2-15 or more years characterized by low but steady viral loads. CD4 counts have been viewed as the best predictor of the risk of developing AIDS-related complications. In the past few years, CD4 counts have been used to monitor disease progression, determine prognosis, select patients for therapeutic trials and monitor therapy. CDC classification of stages of HIV disease is in fact based on CD4 counts and clinical symptoms (1). At the present time, despite its value as a general marker of disease stage, the CD4 counts alone seems to be inadequate to measure prognosis and anti-retroviral therapy (2). In fact, decrease of CD4 counts occur as a result of viral replication and in that sense represent a clinical endpoint rather than a surrogate marker of disease activity. The development of new molecular techniques designed to detect circulating HIV RNA in plasma has allowed a relatively easy quantification of HIV viral load. Viral load and CD4 counts have very different significance. Viral load is a relative measure of virus replication and cell killing, while CD4 count is a measure of the level of immunodeficiency caused by the virus. Therefore CD4 counts remains an essential index for making decisions regarding prophylaxis for opportunistic infections and for evaluating the immunological effects of antiretroviral therapy.

 2. T LYMPHOCYTE SUBSET ALTERATIONS

During the course of HIV disease CD4 T cell counts decrease, while CD8 T counts are elevated and CD3 total T counts remain constant (3). A role for homeostatic control of the total T cell number irrespective of CD4 and CD8 subset has been proposed (4). The rate of loss of CD4 cell varies among patients. In general, CD4 decline occurs in four stages (3): 1) a rapid decline for 12-18 months at the time of seroconversion, when CD4 drops from a median of 1000/mm3 o 50% to 600/mm3 o 30%, 2) a plateau or gradual decline which can last several years during the asymptomatic period (CD4 median 500/mm3, 25%), 3) a more rapid decline during the several months just before AIDS develops. This, generally, takes places when CD4 levels has reached 5-10% or 100-200/mm3, 4) continued CD4 decline until death. Low CD4 cell percentages (<10%, number <100/mm3 and CD4/CD8 T ratio (<0.2) are highly predictive for death from AIDS complications (3). However, although a CD4 count < 200/mm3 is included as an AIDS-defining condition by CDC, there are patients who continue to remain clinically healthy for long time periods, even in the absence of treatment (5). This indicates the importance of host and viral factors, typical of each individual.

CD8 T cell numbers are elevated in HIV-infected subjects. Most of the CD8 cell increase occurs during the first few months of HIV infection and CD8 cell numbers remain relatively stable thereafter (median CD8 value in the asymptomatic phase is 1000/mm3). When AIDS develops, the absolute lymphocyte counts drops down and both CD4, CD8 and CD3 cells decrease in absolute numbers. However the % CD8+T cells increase and in the stage of AIDS almost all T cells are CD8+ (3).

Through the period of HIV infection until AIDS develops total leukocyte count, lymphocyte percentage, absolute lymphocyte number and total T cell levels are comparable to normal controls . In the stage of AIDS, the leukocyte and lymphocyte numbers begin to fall and panleukopenia, lymphopenia and sometimes abnormally low CD3 values of 30%-50% together with a relative increase in % NK cells characterize the final stage of HIV disease (3). In some AIDS patients there are very high levels (>10%) of gd T cell receptor bearing cells (6). In addition, patients infected with Pneumocystis carinii pneumonia show a significant increase in circulating TCS-d1 T cell receptor with respect to asymptomatic HIV-1 seropositive individuals and healthy blood donors (7).

3. NK AND B LYMPHOCYTES ALTERATIONS

There are four levels of alterations in the NK cells of HIV infected subjects (8). These alterations may contribute to the decreased NK cell functional activity throughout the course of HIV disease. First the number of CD16+CD56+ cells, the NK subset with the most potent cytotoxic activity, is markedly reduced. Second, the level of CD16+ expression is decreased, third there is an expansion in the number of CD16dim CD56- cells. Finally, low functional activity of CD16+CD56+ cells in late HIV disease may also occur. These changes in the phenotypes of circulating NK cells in HIV disease have been also observed in donors with CD4+ cell levels >800/mm3 and in donors who had been HIV infected for only 1 or 2 years. The profound NK cell function deficiency in AIDS are believed to contribute to lower host defence against opportunistic viral infections and neoplasias.

B cells can be increased as a result of HIV-induced polyclonal activation and should always be evaluated in children with suspected HIV infection.

The prognostic value of NK and B cell counts has not been established in HIV disease.

4. THE PROGNOSTIC VALUE OF CD4 COUNTS

It is widely documented that even when considering a single count measured at some arbitrary baseline, the CD4 count predicts the development of clinical disease. The risk of developing HIV disease or dying over the next 24 months is 5% among individuals with CD4 counts above 500 cells/mm3 and > 70% among those having fewer than 50 cells/mm3 (2). It is however acknowledged that some patients with high CD4 counts develop AIDS quickly and some with low counts remains disease-free for long periods of time and it is easy in retrospect to conclude that the CD4 count was not a good predictor in these individuals. It is important to identify these individuals in situations in which the outcome measure is survival time (i.e. treatment trials)(9).

5. CD4 COUNTS AS IMMUNOLOGICAL MARKER TO STAGE HIV INFECTION, TO MONITOR DISEASE PROGRESSION AND FOR THERAPEUTIC INTERVENTION

Since CD4 cell counts strongly correlate with stage of HIV-1 disease, a single CD4 measurement is often used to define and inform patients on the stage of HIV infection (1).

To monitor disease progression, CD4 T lymphocyte counts are tested at 6-12 month intervals in asymptomatic HIV infected subjects who are not in therapeutic trials to determine the rate of CD4 cell loss and the level of immunodeficiency. Patients on therapeutic trials are usually evaluated at three months interval or more often during the initial period of treatment. A substantial decline of CD4 percentage, defined as drop of one third or more of CD4 from the previous value represent significant disease progression. A change in the absolute CD4 number in the face of stable CD4 percentages is generally not of clinical significance and mostly reflects biological fluctuations in the absolute lymphocyte counts (i.e. cell migration, circadian rhythms)(3).

The CD4 count is also used as the basis of important therapy guidelines, namely a) initiating antiretroviral therapy when CD4 count is less than 500/mm3 and b) initiating Pneumocystis carinii pneumonia prophylaxis when the measured CD4 counts is less than 200/mm3.

There is evidence that the relationship between immune cell presence in peripheral blood and immune function is at best tenuous, thus suggesting that functional immune impairment may exist long before enumerative impairments are detected.
 

6. SUBSETS OF CD4+T LYMPHOCYTES

Discrepant results have been reported in the literature relative to the loss of the CD45RA subset or the subset defined by CD29 and/or CD45RO. Although some authors originally suggested that there is a preferential loss of one or another of these subsets it was generally accepted that various CD4 subsets were lost concurrently.

Recent studies, however, have suggested that the decline in CD4+ lymphocyte counts during HIV infection is primarily caused by a decrease in numbers of naive CD4+ lymphocytes expressing the CD45RA antigen (10). This subset has a marked increase in level of oxidized gluthatione and decreased ratio of reduced to total gluthatione. This preferential loss is most prominent in advanced clinical and immunological disease, correlated with decreased proportion of CD45RA+CD4+ lymphocytes and may have important immunological consequences including inability to mount responses to novel antigens i.e. opportunistic pathogens as well as the continuously mutating HIV-1 itself. The altered distribution of naive and memory T cell may be involved in the disregulated cytokine production observed in HIV infected individuals (11).

7. ABSOLUTE CD4 NUMBERS VERSUS CD4 PERCENTAGES

Absolute CD4 numbers and CD4 percentages are measurements of CD4 T lymphocyte levels commonly used to define the degree of immunologic impairment induced by HIV (12). The absolute CD4 number is preferred to CD4 percentages to define immunodeficiency. However CD4 percentage is less subject to variation on repeated measurements than absolute numbers (13) since the absolute CD4 number requires three different measurements: % CD4 determined by flow cytometry, WBC and differential counts from hematology counters.

CDC 1993 revised classification system for HIV infection (1) emphasize the use of CD4 counts but allows for the use of CD4 percentages. Equivalences for absolute numbers of CD4+ T lymphocytes and T CD4+ percentages were derived from analyses of more than 15,500 lymphocyte subset determinations from seven different centers, involved in proficiency testing program for lymphocyte subset determinations. The established concordances are : 1) ?500/mm3 concorded with ?29%, 2) 200-499/mm3 concorded with 14-28%, 3) <200/mm3 concorded with <14%.

It is important to keep in mind that these suggested equivalences may not always correspond with values observed in individual patients.

8. MARKERS OF ACTIVATION

CD4 depletion and immune activation represents two different aspects of HIV pathogenesis (14). Since HIV is not completely eliminated from the body by the immune response that takes places during primary infection, its persistence over time will induce a chronic activation of the immune system. This reflects the continuous activation of different effector cells involved in the ongoing immune response to HIV and/or other pathogens.

For these reasons high levels of expression of activation markers are not well correlated with measurements of CD4 T cells.

Total CD8 counts are not predictive for the development of HIV disease because most of the increase in the CD8 subset occurs at the time of seroconversion. On the contrary the presence of the activation marker CD38 is predictive of disease progression. About 20% of the CD8+ cells in healthy subjects are CD38+. There is a progressive increase in the proportion of CD38+(activated and immature) CD8+ throughout the course of the disease from the time of seroconversion. While in other viral infections (CMV, EBV) CD38+CD8+ subset decrease within 1 month from acute infection, in HIV the CD38+CD8+ lymphocytosis never ablates reflecting persistent immune stimulation by HIV. A level > 50% CD38+CD8+ cells is an extremely poor prognostic sign in HIV infected people (15). The HLA DR molecule is expressed on both CD38+ and CD38- CD8 cells. In patients stage CDC II and IV cells HLADR-CD38+ predominate over the other CD8 subsets CD38-DR-, CD38-DR+, CD38+DR+(16). CD8 activation is also observed in the increased expression of CD8 membrane antigen. Moreover, CD8+CD38+DR+ CD8 cells have a higher cytotoxic activity than other CD8+ T cells. CD38 can also be interpreted as a marker of immaturity suggestive of a rapid turnover of these cells.

CD28, the receptor of costimulatory signals delivered by antigen presenting cells through CD80 during T cell activation, is costitutively expressed by more than 50% of normal CD8+ lymphocytes. Primary infection with HIV is characterized by an expansion of CD8+ cells (mean 73% vs 26% in controls), the majority of which (92%) are HLA-DR+. This is mainly due to the increase of CD8+CD28- subset (mean 49% vs 10% in controls ); 66% of CD8+ cells were CD28- compared to 41% of healthy controls (17). The CD28-T cell population is unresponsive to mitogens and it has been suggested to represent terminally differentiated cytotoxic effector cells developing from CD28+ cells after continuous in vivo activation.

HLADR antigen has been shown to be present on a large proportion of CD16+ NK cells from HIV+ patients (18). However this expression does not correlates with the expression of the CD25 antigen resulting in mostly the CD16+/DR+/CD25- phenotype. Activated NK cells shed CD16 antigen (FcgRIII) in the plasma as a result of their activation.

Increased expression of CD20 activation antigen on B cells of HIV individuals has been demonstrated and it is suggestive of polyclonal B cell activation (19). B cells of some HIV infected individuals up-regulate also the expression of FcgRIIb receptors becoming prone to nonspecifically bind antibody aggregates.

9. TECHNICAL ISSUES

The testing process to evaluate CD4+ cells by flow cytometry involves multiple sources of variation (specimen collection, transportation and storage, sample preparation, staining, measurements, data analysis and reporting) that may alter test results. Therefore specific recommendation have been released by CDC (20) and NIH NIAID (21, 22) to standardize methods to reduce the impact of these variables.

9.1 Gating strategies and antibody panels

Monoclonal antibody panels must: 1) allow to determine quality of lymphocyte gate, 2) identify CD4 and CD8 T cells (CD4+CD3+ and CD8+CD3+). CD4 and CD8 are not restricted to the T lineage, CD4 being expressed on monocytes and CD8 being expressed on at least one-third of NK cells. Therefore, accurate measurements of CD4 and CD8 T cells and CD4/CD8 T cell ratio can be determined by the association of CD4 and CD8 with CD3.

Gating is the procedure to identify the cell population to be studied. Gating stategies include: a) the light scatter gating, i.e. gating on lymphocytes selected on FSC/SSC bivariate dot plot, that is used in the 2-color panel (Table I). It relies on "clean" population by scatter and on the validation of the gate by CD45/CD14 immunofluorescence analysis. Gate is established in tube 1 and left unchanged for the analysis of tubes 2-5 , b) CD45/SSC or CD3/SSC, CD19/SSC, CD16/SSC gating strategies that are used in 3-color panels (Table II) and allow the establishment of gating in each tube. Proposed 2-color and 3-color panels are briefly reported here.

9.2 Two-color panel

In 1994 CDC recommended a two-color (mAb FITC-fluorescein isothiocyanate/mAb PE-phycoerithrine) 6 tubes panel (Table I). The panel includes a tube containing isotype control mabs (tube 6).

However this analysis may only be useful to discriminate CD16+ cells from the negative population because of the low CD16 expression, while It is not needed for CD4, CD8, CD3 and CD19 analyses. In addition, in many cases the isotype control may not be optimal for controlling non-specific staining, due to the difference F/P ratio, antibody concentration between the isotypic control and the test reagents.

The complete antibody panel assures the quality of the results accounting for all lymphocytes in the sample, correlate CD4+CD3+and CD8+CD3+results and monitor tube to tube variability in repeated CD3 testing.

CD45/CD14 containing tube allows to establish the proportion of all lymphocytes present in the specimen which are contained within the boundaries of the lymphocyte gate (recovery) and the presence of non-lymphocyte elements present within the gate (purity). Recovery should be at least 90% of all lymphocytes in the sample, purity should be at least 85%.

If the sum % T plus % B plus % NK cells confirms gate purity, correction of subset percentages for lymphocyte purity is appropriate and recommended. (Raw CD4% =41%, gate purity =92%, corrected CD4% = 41%/0.92=45%). For this reason, CD4 evaluation using this mAb panel has to be determined in the context of a complete immunophenotypic profile.

This mAb panel should be used to evaluate new HIV infected patients to rule out other cellular immune abnormalities. A shorter panel including mAb combinations 1-4 can be subsequently used to evaluate HIV-infected adults. The greater danger in using an abbreviated panels is that the internal controls are not longer included.

The major limitation of this panel is that the gate established during analysis of tube 1, although validated with CD45/CD14 staining, does not necessarily include all lymphocyte populations when the sample is stained with other mAb combinations (tubes 2-5). In fact the assumption that gating region characteristics established for one tube can apply to all the other tubes may not be valid and it may lead to systematic errors.

9.3 Three-color panels

Three-color analyses allow the correlation between three cellular antigens and since use cellular markers for identifying lymphocytes the gate can be validated in each tube. For this reason it is not necessary to validate CD4 measurements by identifying 100% of lymphocyte in the gate. Sine analysis of B and NK cells is not necessary, unless they are specifically requested, 3-color immunophenotyping of CD4+ lymphocytes require fewer tubes than 2-color panels.

These panels are particularly useful when scatter characteristics of cells are not good especially when old samples have to be analyzed (23).

Panels I and III use the lineage gating strategy, i.e. gating on bivariate plots such as CD3/SSC, CD19/SSC, CD16/SSC. The selection of CD3+ allows the determination of CD4 and CD8 cells as percentages of CD3+ cells (24). To refer CD4 and CD8 to all the lymphocytes it is necessary to rely on FSC/SSC gate selection and its validation with CD45/CD14 immunofluorescence (panel I Table II). Panel III instead, is limited to Cytoron absolute cytometer with immunocount reagent and must be used in conjunction with absolute counting on the flow cytometer (25).

The NIAID/DAIDS (22) recommends the 3-color panel that uses CD45/SSC gating (Table III), while CDC has not yet released any revised version of the 1994 guidelines.

While for the determination of CD4 T cells only 2 tubes are recommended, B cell determinations are required for paediatric specimens and NK enumeration is necessary to measure total lymphocyte (T+B+NK) recovery.

9.4 Switching from 2-color to 3-color immunophenotyping

As stated in NIAID/DAIDS guidelines, laboratory seeking implementation of any 3-color panel has to demonstrate equivalence of CD4+CD3+ and CD8+CD3+ values measured by the proposed 3-color method as compared to the laboratory’s current 2-color method on a minimum of 60 different specimens with CD4+CD3+ £ 30%.

10. Quality Control

Quality control issues are of concern because flow cytometric data are used to make clinical decisions in the management of HIV infected subjects and appropriate quality control (QC) is critical to assuring that accurate and clinically relevant comparisons can be made at multiple levels: cell to cell within a sample, sample to sample within a laboratory, day to day for a patient and laboratory to laboratory (26).

QC has to be a consistent procedure of the laboratory involved in immunophenotyping. The ‘internal" QC is made of different components : instrumentation, sample preparation and data analysis.

10.1 Instrumentation QC

QC instrumentation involves alignment verification, positioning and anchoring the window of analysis (position of sample space analyzed by the instrument) with the use of stable microspheres (Flow Cytometry Standards Corp., FCSC). Positioning results from PMT adjustments and anchoring can be achieved by channel targeting. Another problem is compensation for spectral overlap that must be corrected for the accurate quantitation of cells labelled with fluorochromes. Care should be taken that the spectral emission overlap is not overcompensated (27). Overcompensation will result in the erroneous loss of doubly-labelled cells. In order to avoid this risk, compensation must be adjusted using representative stained cell suspensions containing negative and positive populations rather than a mixture of blank and stained microbeads. Compensation the fluorescence level of stained microbeads to that of blank microbeads still will result in some degree of overcompensation, since blank beads are representative for the instrument’s noise level, which typically is below the level of cellular autofluorescence. Therefore, to ensure accuracy, it is recommended that compensation be checked with control cells that are labelled with the actual antibodies used in the assay.

Maintenance of appropriate instrument settings is performed by assessing tolerance limits for each parameter by repeated measurements of calibration standards tagged with the relevant fluorochromes (28). Tolerance limits for FSC and SSC are assessed using cells while tolerance limits for fluorescence are assessed by running the fluorescent calibration standards. The instrument settings can then be monitored over prolonged periods of time by checking whether or not measurements of calibration standards after each cold start yields results within the tolerance limits. If not adjustment of instrument settings is required. In case a fluorescence parameter is involved, adjustment of the correction for spectral overlap is also necessary.

10.2 Sample QC

Sample or biological QC involves establishing the adequacy of sample preparation and staining procedures. Where light scatter patterns are used to discriminate, conditions that affect light scatter resolution (lysing, fixation techniques) must be optimized. Where low intensity signals must be resolved , the selection of fluorochrome or amplification technique must be optimized to give adequate signal/noise. Where precise measurements of relative signal intensity are required, reagent stoichiometry and preparation effects on labelling efficiency must be optimized (26). The type of quality control specimen to be used is a normal blood from a healthy individual, prepared using the laboratory’s standard protocol. Other type of quality control specimens (Cyto-trol, CD chex) are commercially available. The value of this materials is that they provide a more constant reference point than do normal specimens from different individuals. Their limitation is that they sometimes cannot be used with the same sample preparation procedures as test samples and their stability is limited.

The pH of the medium or of the fixative is very important to the resulting fluorescence intensity of the sample, for fluorochromes that require ionization to fluoresce, i.e. fluorescein. The longest-wavelength excitation and emission peak heights are considerably diminished at acidic pH. Carefully controlled pH conditions are required when performing quantitative fluorescence intensity measurements.

Negative controls are used to define the background signal found when there is believed to be no specific biological signal of interest, whereas positive controls are used to established that the signal above background can in fact be detected for a specimen known to exhibit a particular biology or pathology (26).

11. Analysis QC.

Gating assumptions and data analysis methods should be considered as potential sources of variability and the goal is to identify controls and monitor critical variables that can bias the results. Gating strategies may be either inclusive, to maximize the inclusion of relevant cells or exclusive to minimize the inclusion of contaminating cells (26). Gating parameters can be either intrinsic (light scatter) or extrinsic (fluorescence) or a combination of the two. In situations where a pair of gating parameters fall into one of the two categories, the gating is considered homogeneous (FSC/SSC). If gating criteria use signals from both categories, then the gating is considered heterogeneous (CD45/SSC).

When immunofluorescence CD45/CD14 is used as an aid to define the optimal light scatter gate for lymphocyte acquisition, the procedure is referred to as "back gating". This procedure can be useful also with other mAb combination to define physical parameters of ungated stained cells .

12. Data reporting QC.

There are several checks on immunophenotyping results to determine whether flow results are accurate (15). First the lymphosum or sum of T+B+NK should be in the range 100% ± 5% after each of these values is corrected for the percentages of lymphocytes in the lymphocyte gate. If the lymphosum is 85% or 120% the cytometrist should check for cursor setting errors, light scatter distortion, improper gating or staining errors. Either all or the selected samples from the panel can be restained and rerun to validate the results. A second useful check is whether the sum of CD4 + CD8 T cells approaches the CD3+ cells within 10% of the CD3 value, the missing cells being CD3+4-8- gd cells.

13. Quality assessment programs.

Quality assessment programs are used to evaluate proficiency in determining lymphocyte phenotype and are based upon shipments of whole blood specimens or control cells to several laboratories. Comparative studies determine interlaboratory variation i.e. homogeneity of reported data and the accuracy of data of each laboratory. Accuracy is generally evaluated as the difference between the reported value and the median value of the distribution of all data (residual value). The median value is taken as the "true" value for a particular parameter (i.e. CD4+ cells).

Certification of laboratories that participate to the program designed by NIAID DAIDS in USA (21) are based on the following criteria. The unsatisfactorily level of performance is defined as 33% of a laboratory's CD4 analyses with residual values greater than or equal to ±5% and with deviates greater than or equal to ± 2. The deviate is the residual value divided by the interquartile range (IQR= 75th - 25th percentiles of participating laboratories).

In Italy the Liguria Region QC program adopted these criteria since 1992 and proposed a new criterion for absolute counts (acceptable residual being ±100/mm3 CD4 cells). These criteria are currently used also in the National QC program for CD4 evaluation promoted by the Italian Health Institute, Ministry of Health. Laboratories identified outside the acceptable range of performances are encouraged to improve their results and to conform to the other laboratories.

In general, quality assessment programs are used as the basis for the accreditation of laboratories involved in lymphocyte immunophenotyping.

2. Saag M.S., Holodniy M., Kuritzkes D.R., O'Brien W.A., Coombs R., Poscher M.E., Jacobsen D.M., Shaw G.M., Richman D.D., Volberding P.A. HIV viral load markers in clinical practice. Nature Medicine 1996. 2: 625.

3. Giorgi J. Characterization of T lymphocyte subset alterations by flow cytometry in HIV disease. Ann. NY Acad Sci. Clinical flow cytometry. 1993. 677: 126.

4. Adleman A.M., Wosfy D. T-cell homeostasis: implications in HIV infection. J. Acquir. Immune Defic. Syndr. 1993. 6:144.

5. Hoover D.R., Rinaldo C., He Y., Phair J., Fahey J., Graham N.M.H. Long-term survival without clinical AIDS after CD4+ cell counts fall below 200 x 106/L. AIDS 1995. 9: 145.

6. Margolik J.B., Scott E.R., Odaka N., Saah A. Flow cytometric analysis of gd T cells and natural killer cells in HIV infection. Clin. Immunol. Immunopathol. 1991. 58: 126.

7. Agostini C., Zambello R., Trentin L., Semenzato G. T lymphocytes with gd T-cell receptors in patients with AIDS and Pneumocystis carinii pneumonia. AIDS 1995. 9: 203.

8. Hu P., Hultin L.E., Hultin P., Hausner M.A., Hirji K., Jewett A., Bonavida B., Detels R., Giorgi J.V. Natural killer cell immunodeficiency in HIV isease is manifest by profoundly decreased numbers of CD16+CD56+ cells and expansion of a population of CD16 dim CD56- cells with low lytic activity. J. Acq. Immune Defic. Synd. 1995. 10: 331.

9. Sabin C.A., Mocroft A., Phillips A.N. The use of CD4 counts as prognostic markers in HIV infection. AIDS 1995. 9:1205.

10. Aukrust P., Svardal A.M., Müller F., Lunden B., Nordøy, Frøland S.F. Markedly disturbed glutathione redox status in CD45RA+CD4+ lymphocytes in human immunodeficiency virus type 1 infection is associated with selective depletion of this lymphocyte subset. Blood 1996. 88: 2626.

11. Fauci A.S. Multifactorial nature of human immunodeficiency virus disease: implications for therapy. Science 1993. 262:1011

12. Taylor J.M., Fahey J.L., Detels R., Giorgi J.V. CD4 percentage, CD4 number, and CD4: CD8 ratio in HIV infection: which to choose and how to use. J. Acq. Immune Defic. Synd. 1989. 2:112.

13. Malone J.L., Simms T.E., Gray G.C., Wagner K.F., Burge J.R., Burke D. S. Sources of variability in repeated T-helper lymphocyte counts from human immunodeficiency virus type-1 infected patients: total lymphocyte counts fluctuations and diurnal cycle are important. J. Acq. Immune Defic. Synd. 1990. 3:144.

14. Pantaleo G., Fauci A.S. New concepts in the immunopathogenesis of HIV infection. Ann. Rev. Immunol. 1995 . 13:487.

15. Landay A., Ohlsson-Willhelm B., Giorgi J.V.. Application of flow cytometry to the study of HIV infection. AIDS 1990; 4:479.

16. Ramzaoui S., Jouen-Beades F., Gilbert D., Borsa-Lebas F., Michel Y., Humbert G., Tron F. During HIV infection, CD4+CD38+ T-cells are the predominant circulating CD4+ subset whose HLA-DR positivity increases with disease progression and whose Vb repertoire is similat to that of CD4+CD38- T-cells. Clinical Immunol. Immunopathol. 1995. 77:33.

17. Brugnoni D., Prati E., Malacarne F., Gorla R., Airò P., Cattaneo R. The primary response to HIV infection is characterized by an expansion of activated CD8+ CD28- cells. AIDS 1996. 10: 104.

18. Lucia B., Jennings C., Cauda R., Ortona L., Landay A. Evidence of a selective depletion of a CD16+ CD56+CD8+ Natural killer cell subset during HIV infection. Cytometry (Communications in Clinical Cytometry) 1995. 22:10-15.

19. Staal F.J.T., Roeder M., Bup J., Mole L.A., Anderson M.T., Rajui P.A., Israelski D.M., Deresinski S.C., Moore W.A., Herzenberg L.A., Herzenberg L.A. CD20 expression is increased on B lymphocytes from HIV-infected individuals. J. Acquir. Immune Defic. Syndr. 1992. 5: 627.

20. Centers for Disease Control. 1994 Revised guidelines for the performance of CD4+ T cell determinations in persons with human immunodeficiency virus (HIV) infection. MMWR 1994. 43 (NoRR-3): 1-21.

21. National Institute of Allergy and Infectious Diseases (NIAID), Division of AIDS, Flow Cytomtery Advisory Committee. 1993 Guidelines for flow cytometric Immunophenotyping.

22. Nicholson J.K.A., Kidd P., Mandy F., Livnat D., Kagan J. Three-color supplement to the NIAID DAIDS guideline for flow cytometric immunophenotyping. Cytometry (Communications in Clinical Cytometry) 1996. 26:227.

23. Nicholson J.K.A., Jones B.M. & Hubbard M. CD4 T-lymphocyte determinations on whole blood specimens using a single-tube three-color assay. Cytometry 1993. 14: 685.

24. Mandy F.F., Bergeron M., Recktenwald D., Izaguirre C.A. A simultaneous three-color T cell subsets analysis with single laser flow cytometers using T cell gating protocol. Comparison with conventional two-color immunophenotyping method. J. Immunol. Methods. 1992. 156: 151.

25. Mercolino T.J., Connelly M.C., Meyer E.J., Knight M.D., Parker J.W., Stelzer G.T., De Chirico G. Immunologic differentiation of absolute lymphocyte count with an integrated flow cytometric system: a new concept for absolute T cell subset determintations. Cytometry (Communications in clinical cytometry). 1995. 22:48.

26. Muirhead K.A. Establishment of Quality Control procedures in clinical flow cytometry.
Ann. NY Acad Sci. Clinical flow cytometry. 1993. 677: 1.

27. Schwartz A., Fernandes-Repollet E. Development of clinaical standards for flow cytometry. Ann. NY Acad Sci. 1993. 677: 28.

28. Flow Cytometry Working party in the Netherlands (SIHON/ SKMI/ SKZL). Standard operating procedures: setting up, calibration and adjustment of a flow cytometer for multicolor immunophenotyping. 1996.

29. Shapiro H.M. Practical Flow cytometry: parameters and probes. New York, Liss, 1994. 29.