Comprehensive & Cost Effective Laboratory Monitoring of HIV/AIDS: an African Role Model

Introduction

The modern scientific efforts in the area of biomedicine lead to practical benefits including the introduction of efficient diagnostics for infectious diseases. High quality and fast diagnostics are needed because of the danger of disease spread due to global travel and the current devastating impact of viral and bacterial infections in the poverty stricken areas of the globe¹.

In resource poor countries treatment of HIV positive patients with anti-retroviral therapy (ART) has been successfully introduced^2,3. The protocols of ART are currently based on (i) the serological tests for HIV+ individuals, (ii) the once-daily combinations of genetic drug regimens used as 'first-line' ART and (iii) the CD4 and/or viral load assays to decide when to start ART. The protocols are, however, changing as the scope for ART can now be further widened with more relaxed guidelines to initiate ART earlier at higher CD4 counts ². Importantly, increasing the number of HIV+ patients who receive ART at higher CD4+ cell/μl of blood may even slow down the spread of HIV despite the current lack of HIV vaccine ^{4, 5}.

Thus HIV infection is poised to become a chronic disease in increasing numbers of survivors, requiring organized medicine more than ever but the poor state of laboratory medicine in Africa may still remain a barrier of effective health care ⁶. There are voices to emphasize that the provision of short-term grants does not provide sustainable solutions ^7-9, and the HIV-related grants supplied by Western countries at the current scale cannot continue forever ^8,9. In 2008 USD 15.6 billion was spent on AIDS programs in low and middle income countries, and about 50% of these were from external, including North American, sources. The calculated trends show that this expenditure is to increase up to 2-3 times higher level, and that the epidemic will still be with us in 2031 unless there is a change in the current overall approach ⁹.

Thus African countries must contribute by better HIV prevention and wiser cost-savings ⁹. Here we suggest that the most recent inventive 'role models' for routine work will be established and introduced in resource-poor countries (e.g. in Africa or India). These are the areas of the globe where the scientists will learn to fully benefit from the current international support and then to utilize their skills to further improve every-day technology, by contributing to a global development.

In western medicine the diagnostic activities in large academic centres and hospitals are frequently "departmentalized", i.e. separate "zones of interest" may exist for the different disciplines leading to a fragmentation into haemato-, viro-, retroviro-, bacterio-, sero-, immuno-, and cytological laboratories instead of conducting an unifying approach using a minimal set of multidisciplinary diagnostic assays in an infectious-disease laboratory. In addition, the diagnostic tests can also be far too complicated and costly. Clearly, such services would need to be better coordinated but these unifying efforts are unlikely to come entirely from western directions. Consequently, the leading laboratories and institutes, such as the National Health Laboratory Service (NHLS-SA) in cooperation with University of Witwatersrand in Johannesburg, South Africa, are facing an the extra task to further develop, modify, re-focus and improve their cost-effective diagnostics based on modern technology.

In the current video we describe flow-cytometric methods to monitor the anti-retroviral therapy (ART) in HIV-infected patients. This technology is presented step-wise in six modules for a combination of relevant haemato-, immuno-, and virological assays in a single tube, combined with essential internal quality control (QC) to ensure reproducibility of testing ¹⁰. The uses of monoclonal antibodies employed here are all scientifically fully established. The CD45 reagent is for white blood cell counts ^11-13, combined with CD4 to enumerate the "helper" T-lymphocytes which are the primary targets for HIV ^12-14. It is a "constructive trick" that in the same tube the CD38 activation marker is detected on CD8+ T cells as a HIV viral-load (VL) related assay ^15-17. In order to emphasize the South-African contribution, we point out that these concepts and their cost-effective practical combinations have all been worked out in NHLS-Wits collaborations based on leading western technologies ^{10,12,13,16,17}.

These assays, carried out in combination, are going to be needed for the proper monitoring of patients who are undergoing long-term ART. When the cost-efficiency is scrutinized, this coordinated flow-cytometry technology is proving to be economical ^16,18,19. These single-tube assays represent cost-saving over their western variants and over the similar tests done by separate instruments and/or carried out as different surrogate tests using other methods. These savings are itemized in the discussion.

For the acceptance of this basically simply technology, which utilizes four-color flow cytometry in a single tube, it is essential to show the technical steps in a video format. Expert advisors, laboratory managers and policy makers who carry the duty of making decisions about introducing modern medical technology are frequently not in a position to see the latest technical details as carried out in the large regional laboratories with huge burdens of work load. Hence this video shows details of new developments that have introduced in South Africa to optimize the workflow for testing >1000 clinical samples each working day ^10,13,16,17.

Protocol

Flow cytometry

The tests were carried out by 4-color flow cytometry but the assays can also be performed using 2-colors in two parallel tubes (see below). The Beckman-Coulter XL-MCL, a cytometer of stable fluidics ²¹, is used but the same technology may also be applicable to other cytometers. The technology described handles large daily workloads of ~800-1100 samples/day^10,13.

Module 1: Set up of instruments

Daily in-house maintenance is cleaning perform twice each day: on start-up and on completion of the last patient sample. This includes the flushing of vacuum lines and acquiring Clenz (5% sodium hypochlorite; BC, Miami, FL). Also, at the end of each 30-tube carousel Clenz and distilled water is run through the system to minimize protein build-up.

At daily start-up the vacuum- and pressure readings are checked. It is confirmed that the background counts remain acceptable (<100 events in 100 sec). After running the first blood sample it is established that in an aliquot of sheath fluid the carryover is <1%. Then laser alignment and fluorescence sensitivity are monitored daily using a sample of FlowCheck ²¹. It is observed that the Mean Channel Numbers are consistent (500±50) with a low variation (half peak CV values <2%). The compensation is checked weekly; FlowSet beads are plotted to see whether the mean channel values are within 1% of the preset target values ²². To visually confirm the precision on Levey-Jennings plots, the mean channel numbers and half peak coefficient of variation (HPCV) values for the four Fluorescence channels (FL1-FL4) are daily displayed. Next the accuracy of CD4 counts is confirmed at the two CD4 levels. During a PLG-CD4 test stabilized blood process controls (IMMUNOTROL) are used the low normal (450-700 CD4 cells/μl) and the low levels (100-200 CD4 cells/μl, depending on batch; ref 23). These Immunotrol values are required to be within 0.5 SD of the mean with a %CV of < 5% as requested by the international BC-IQAP program ²⁴, UK NEQAS ²⁵ and AFREQAS ²⁶. It is important that with these simple organized measures as shown in the video the robust flow cytometers can maintain the required stability for both absolute cell counting such as CD4 tests and for fluorescence intensity measurements used for the activation marker CD38. This stability has recently been documented in detail ¹⁷ confirming previous studies using other cytometers ^15,27.

During weekly maintenance the list mode data (LMD) files are archived for future reference and this archived material is then deleted from the cytometer hard drive.

Monthly maintenance includes the cleaning of reagent containers, the sheath, clenz and waste tanks. The carousel-mode sample head is cleaned while all filters are also removed and washed. This cleaning is also extended to preparation stations TQPrep and PrepPlus2 (BC, FL). The pipettes used for manual preparation are validated in-house at least every 3 months but if pipetting inaccuracy is suspected the pipette is immediately taken out of service for corrective action. Six monthly maintenance is scheduled by the Beckman Coulter, SA.

Module 2: Preparing samples

For the CD45/CD4 two-color PanLeucogated protocol ⁵ the Beckman Coulter FlowCARE kit ²⁸ is used in a lyse-no-wash procedure (IMMUNOPREP, Beckman Coulter, FL). This kit contains CD45-fluorescein-isothiocyanate (FITC; green) and CD4-phycoerythrin (PE; orange). Flow-Count beads²⁹ are purchased in bulk and are added to play a special operator-friendly role as known numbers of beads in a known volume are pipetted on the spot into the sample ³⁰. This is done with the same pipetting as used to deliver the volume of blood. Thus the Bead Count Rate (BCR) in the expected range shows that the pipetting in the given sample is precisely done, as described in Module 3 below. Importantly, here the simple pipetting method, referred to as 'forward pipetting', works best, providing that the bead- and blood-volumes are delivered with the same pipette and tip and with identical handling 30 . We found that forward pipetting yielded optimal overall uniformity between laboratories, both with respect to operators achieving pipetting targets within acceptable tolerance limits and reproducibility of repeated dispensing as measured by the co-efficient of variation (%CV) ^10,31,32. This method, presented at local congresses ³², is approved for use under the South African National Accreditation Scheme (SANAS, www.SANAS.co.za) guidelines. This BCR use of beads is different from the other volumetric counting methods where most frequently the lyophilized bead product such as the BDS Trucount preparations are used which contain known numbers of beads. These preparations are excellent to document the stability of the fluidics of the flow instruments by the Flow Count Rate (FCR) method ^20,31. Nevertheless, the use of ready made bead-containing tubes renders the technology more complicated because here it is essential to deliver the blood volume into these tubes with utmost absolute precision. It is generally agreed that for that purpose only the "reverse-pipetting" technology is acceptable following applicable training. Our BCR method is therefore, less expensive, more informative, simpler and easier to learn -as shown in the video - when compared to the alternative similar methods ^10,31,32.We also find that the operators using forward pipetting are less likely to aspirate sample into the pipette which with reverse pipetting can lead to both the damage to the internal mechanism and the need for re-calibration which is difficult in remote locations. Our BCR method also obviates the need for automated pipetting systems in the small-to-medium sized laboratories (List 1).

The blood samples are always processed in Class II biohazard cabinets. In order to avoid extensive noise during filming, the video shown is a re-enactment with safe materials outside these cabinets. The PLG-CD4 acquisition protocol on the cytometers allows a time to stop at a given time limit or continue to acquire 5000 lymphocyte events to improve accuracy on samples with low CD4 counts.

Module 3: The Bead Count Rate (BCR)

Monitoring bead count rates (BCR) in our system or each control sample is calculated using the formula bead events/seconds ^{10,13,30,32,33}. The Mean, SD and %CV values are used on a spreadsheet to generate a sequence plot for identifying BCR outliers ³⁰. These confirm the optimal operation of the three systems of pipetting, processing and analysis referred to as sample preparation verification (SPV), applied to both the samples pipetted manually or by using the PrepPlus2 automated pipetting instrument ¹⁰.

During acquisition the BCR results are directly viewed in real time on monitors. On the BC cytometers the value of mean BCR is 80±5 events/sec. We show that the aberrant BCR-s are identified and re-run before the patient results are released. The decision-tree algorithm is illustrated in Figure 1 (see also ref.10).

Instrument malfunctions are indicated by fluctuations in BCR when monitored during longitudinal precision verification (LPV) to detect pipetting trends, patterns and sudden shifts¹⁰. Fluidics precision verification (FPV) is an optional in-house procedure where a single dilution of beads is aliquoted into 10 tubes for serial acquisition to confirm signs of fluidic instability on a potentially misbehaving instrument ¹⁰.

Module 4: CD45

For routine absolute counts in our studies a full blood count analyser LH-750 (Beckman Coulter), calibrated monthly, is used daily as a reference for white cell counting (WCC) against which the flow cytometers are compared. The CD45-generated WCC-s on the flow cytometers are also validated daily by comparison with the three levels of stabilized Beckman Coulter 5C whole blood cell controls. The acceptable limits of WCC agreement between the flow CD45 counts and full blood count analysers are <5% ¹⁰. This indicates that when the resource restrictions are more severe then in our service than one single instrument, such as a flow cytometer, seems to be sufficient to provide the clinically appropriate WCC including CD4 counts.

White cell differential counts are performed on various haematological analysers which differ, particularly in assessing monocyte counts. The pattern of differential CD45 expression is a classical use of flow technology for differential WCC ^{11,12, 34}. As the flow cytometer-derived CD45 WCC and differential counts are proven to be accurate, precise and robust, these can be routinely incorporated into the PLG-CD4 test method and also made available to clinicians at no extra cost.

Abnormal CD45 patterns are noted on CD45/side scatter plots of the PLG-CD4 protocol and followed-up on a stained blood film to prompt further clinical investigation. This method detects infection (bacterial and viral), leukaemia and lymphoma on both HIV-negative and HIV+ patients ^{34, 35}. During a routine laboratory follow-up in HIV+ patients the full blood count is frequently not requested from the haematology laboratory. Therefore the abnormal CD45 patterns are diagnostic when candidly observed in samples sent for CD4 counts. An additional advantage is that the accurate absolute CD4 count can still be made, even in patients with findings of leukaemia or lymphoma in the blood ³⁵.

Flow CD45 is well suited to standardize haematology. The haematological instruments significantly differ from each other. We suggest introducing the CD45-based flow cytometric quality control for the WBC differentials to compare and standardize the haematological analysers ^10,36.

Module 5: CD4

The PLG-CD4 Protocol introduced on the Beckman Coulter XL software consists of eight dot-plots. Of these 4 are used to determine the CD4 counts and 4 are for quality monitoring ^12,13. Among these, the two dot-plots shown in Fig.2 are important. All white cells are identified using CD45 and side scatter (region A dot-plot 2) while debris, red blood cells/ghosts and platelets are ignored. The CD45+ events are then transferred to dot-plot 3 where CD4+ lymphocytes are identified by their bright CD4 expression and low side scatter (region C). Absolute CD4+ lymphocyte count are calculated from the number of events in region C and WCC (in the Flow Count calibration region F). The values for CD4% of lymphocytes are obtained from region B dot-plot 2. The strongly CD4+ T lymphocytes and the weakly CD4+ monocytes are discriminated by CD4-fluorescence intensity and scatter features without a need for using an extra CD3 T cell specific antibody in the reagent cocktail (reviewed in ref.13).

Module 6: CD8/CD38

CD8-CD38 staining is a part of the 4-color staining using CD45-FITC/CD4-PE/CD8-ECD/CD38-PC5. The CD4/CD45 is the same as above (Flow Care PLG-CD4), and the two extra reagents added are CD8-ECD/CD38-PC5 ¹⁷. If a cytometer runs only two colors, an extra tube is added to perform CD8-FITC/CD38-PE two-color staining, using again the lyse-no-wash procedure (IMMUNOPREP, Beckman Coulter, FL). A gating strategy similar to CD4 is used to identify CD8+ T cells with CD8-ECD vs. side scatter (Figure 3). CD38 expression on the CD8+ lymphocytes is plotted as single parameter histogram of CD38-PC5 mean fluorescence intensity (MFI) vs. cell count ^16,17. The CD38 MFI values are plotted sequentially for each visit of the same patient. The first visit (prior to treatment) is the 'baseline' MFI value to which every other visit of the same person's CD38 MFI is compared¹⁶.

Controls and instrument stability as described in Module 1 is based on daily acquisition of Flow Check to monitor laser alignment and voltage stability. Flow Set sets the color compensation for all the four fluorescent channels (see above). Immunotrol is the 'negative' process control sample. The 'positive' controls are the monocytes in normal blood expressing 9 x 10³ CD38 molecules per cell, and a known HIV+ sample untreated ¹⁷.

The network of the South African National Health Laboratory Service ( SA-NHLS) laboratories has adopted the

quality control methods standardized across 70 laboratories^{13, 33} with a centralized database containing standard operating procedures and training material. Laboratories in the network are helped by discussions, further training and regular audition including the scrutiny of pipetting technology ^32,33. The quality assessment programme is monitored monthly. In this way, quality management is devolved regionally but high standards are maintained with the contribution of the SA-NHLS QA Department^{13, 26}.

List 1. Potential cost saving with coordinated cytometry in HIV disease:

1. The use of a single instrument with familiarity and service contract, instead of applying multiple haematological, immunological and virological instruments
2. The performance of assays in a single tube, without extra costs and without sending samples to and from different locations including both WCC, WBC differential counts CD4 testing and CD38MFI monitoring to predict response to ART
3. Immediate reporting of response to ART with CD38 MFI reporting, significantly improving turnaround times relative to traditional HIV VL reporting (real-time)
4. The clarity and clinical use of the combined assessment of results on the same reporting forms
5. The need for automatic pipetting instruments may be negated as the simple method of "forward-pipetting" is used and the samples are monitored by BCR,
6. During operation, inaccurate handling of samples is identified and the problematic test results are rectified before results are authorized
7. Proactive preventive management for instrument malfunction leading to early service booking prior to a severe crash
8. Diagnosing unsuspected cases of leukaemia/lymphomas during the routine service activity
9. Extension of similar flow technology for leukaemia phenotyping and monitoring
10. Coordination of the CD38 test with storage of dried blood spots (DBS) for HIV VL testing in all patients and saving costs by sending DBS for evaluation only if there are clinical or CD38 indications to do so
11. Similar flow technology to replace interferon-γ ELISpot method in tuberculosis assays
12. Quality control for peripheral laboratories which perform the "point-of-care" CD4 tests
13. Saving the costs of "unwanted" reagents such as antibodies to CD3 and subset markers (but purchasing CD38 instead)
14. Bulk purchase of inexpensive beads for BCR testing (instead of buying expensive pre-pipetted counting-tubes)
15. In a network, joint negotiations for advantageous bulk prices with the commercial companies

Figure 1. A decision tree/algorithm used when a sample tested shows an outlying result of the bead count rate (BCR). The main task is to exclude pipetting error versus intermittent volumetric failure. The diagram shows the probable causes of errors. It is also indicated that which protocol histograms and statistics are affected ¹⁰.

Figure 2. An example of the PLG CD4 protocol that incorporates CD45-FITC (green; x-axis) vs. side scatter (complexity; y-axis) and CD4-PE; (orange; X-axis) vs. side scatter (complexity; y-axis). The gate 'A' is the total white cell population with the selected gate of lymphocytes 'B'. The gate 'A' is transferred to the second diagram on the right in order to identify the CD4+ lymphocytes in gate 'C'.

Figure 3. The combination of CD45/CD4 panleucogating assay (in diagrams '2' and '3'; see also Fig.2) with the CD8/CD38 assay (in diagrams '4' and '6). When the total leucocytes (gate 'A') are transferred to diagram '4', here the CD8+ lymphocytes are gated on the basis of CD8-ECD ('x'-axis) and side scatter ('y'-axis; as shown in gate 'D'). Then the CD8+ cells are transferred to diagram '6' to visualize the CD38-PC5 staining intensity in a single-parameter diagram. This is shown as the parameter 'G' in diagram '6'. The mean fluorescence CD38-PC5 intensity (MFI) is recorded. On ART this value gradually decreases ^16,17.