Final Report Summary - PREVENTIT (ASSURED Point-of-Care Device for Syphilis and HIV in Pregnant Women and New Born)
Executive Summary:
Mother to child transmission of syphilis and HIV are major causes of severe morbidity and early death of the young infant. Treatment of the mother reduces the risk of congenital transmission. Co-infections are common and most cases of congenital transmission of HIV and syphilis infection take place in hard-to-reach populations, where the availability of an accurate diagnostic test for onsite screening of pregnant women is essential to gain access to and to deliver tailored medical care. The use of a point-of-care test for the combined detection of HIV and syphilis infections in pregnant women would allow same day testing and treatment thus reducing the number of patients that are lost during follow-up and increasing the number of infected persons that are receiving treatment.
In this project the application of two different test platforms for point-of-care testing for HIV and syphilis infections was investigated. The first platform is the lateral flow assay (LFA) while the second test format (Cyclotek) utilizes a platform specifically designed for multi-analyte testing. In this platform, the complex assay strip that forms the heart of the LFA and is composed of a number of overlapping components is replaced by a single moulded solid matrix with a circular groove on which the reaction takes place. The use of a single component matrix simplifies manufacturing and whereas space in LFA for the application of capture probes is limited, multiple capture probes can be immobilized at defined positions in the Cyclotek reaction groove. Moreover, potential for improved sensitivity of the Cyclotek assay compared to LFA was anticipated; the circular design permitting multiple rounds of interaction between serum molecules and probes. Using these two platforms, the multiplexing of HIV-1, HIV-2, Treponemal and non-Treponemal probes for the detection of HIV and syphilis antibody responses were investigated. To improve test performance and standardisation through operator independent automated readout, the use of fluorescent probes was investigated in addition to conventional visual detection.
Options for multiplexing HIV and syphilis capture and colloidal gold-labelled detection probes appeared to be restricted due to the different biophysical properties of the syphilis and highly hydrophobic HIV antigens and the resulting narrow range in assay conditions for the optimal functioning of individual probe combinations. In particular assay conditions required for non-Treponemal (cardiolipin) antibody testing turned out to be difficult to combine effectively. Combining HIV-1 and HIV-2 antibody testing with Treponemal specific antibody in LFA with visual read-out gave a sensitivity of 98.8% for HIV-1, 100% for HIV-2 and 72.7% for syphilis at a specificity of 97.7%, 100% and 90.9%, respectively. The majority of the syphilis positive samples that did not react had low values in the reference EIA and were classified as earlier syphilis after treatment or lues latens based on a negative Venereal Disease Research Laboratory (VDRL) test. The sensitivity for VDRL test positive samples classified as lues latens/after treatment/long existing lues was 84.2% and 100% for active syphilis/recently treated. An independent evaluation to confirm results is being performed and further fine-tuning may improve the diagnostic accuracy of the syphilis components. The development of the Cyclotek system appeared to be even more challenging. Conditions for HIV testing did not match those for syphilis and conditions for combined detection resulted in a reduced signal intensity for HIV with cross-reactivity observed between HIV and syphilis probes resulting in a suboptimal diagnostic performance. The use of fluorescent detection probes did not improve test performance. Of various fluorescent labels tested, including novel fluorescent polymers and a number of well-established commercial benchmark fluorophores Sirigen’s BV605 gave the brightest signals and is a very promising detection label for LFA. However, due to the characteristics of the HIV and Treponemal antigens, the conditions required for combined detection reduced the signal strength of the fluorophore, in turn reducing assay sensitivity and specificity. The use of automated reading adds to standardization, in particular for specimens that give a low or borderline signal intensity, but did not improve diagnostic accuracy.
Standard operation procedures have been prepared for the HIV/Syphilis LFA prototype assay and product manufacturing file highlighting essential steps in the development of the assay is available.
For field evaluation in pregnant women in mother and child clinics in sub-Saharan Africa an evaluation batch was prepared and documents to obtain ethical approval were prepared. Due to an unfortunate management decision of partner 1 the project was ended before the anticipated end date and field studies to confirm the excellent diagnostic performance in the target group and to develop guidelines for best practices could not be performed.
Project Context and Objectives:
Current point-of-care diagnostics for infectious diseases are based mainly on the immunochromatographic lateral flow assay (LFA) format which technology. The number of probes for different pathogens that can be combined in a single LFA is limited by the design of the assay and the physical format of the LFA. Also, in LFA the clinical specimen and detection reagent migrate over the test line once and are allowed to react with the immobilized capture probe during a short period, potentially limiting sensitivity. The Cyclotek platform is specifically designed for multi-analyte detection by using a single solid detection surface that allows the deposition of capture probes at multiple positions. The Cyclotek system may also improve sensitivity as the system is designed for multiple rounds of interactions between the capture probes, serum and detection reagent.
LFA point-of-care tests that utilize coloured detection reagents such as colloidal gold nanoparticles or stained latex particles enable reading of test results by visual inspection. The contrast of the detection reagent limits sensitivity and visual inspection of LFA test results affects reproducibility, especially at the lower end of the detection limit. Reading of test results by visual inspection is greatly influenced by (artificial) lighting conditions and found more difficult by users wearing reading glasses – thus the current trend directs towards the use of readers for standardized LFA test reading and interpretation. The use of fluorescent labels may improve sensitivity and simplify automatic read-out. If linked to an online database the use of readers provides many other advantages that benefits the health system and patient care such as improved patient registration and disease surveillance.
The PreventIt project focuses on the development of a point-of-care test for the serodiagnosis of HIV and syphilis infections as a model for a multi-analyte testing. HIV and syphilis infections during pregnancy pose high risks for the infant. Their timely treatment in pregnant women prevents congenital transmission and reduces the risk of an adverse pregnancy outcome. Co-infection are also very common among other risk groups. Notwithstanding, pregnant women with a high risk of infection do not always receive the required medical attention. Often, if testing is available it is done for only one of the two infections. Thus, the use of point-of-care testing is essential to provide diagnostic services to such hard-to-reach groups, allowing same day testing and thereby reducing the number of infected individuals who do not receive treatment.
A test for HIV infections should include reagents for the detection of antibodies against different HIV types, in particular when used in the African context. The immune response in syphilis involves production of Treponemal antibodies, which are specific for T. pallidum antigens, and non-Treponemal anti-phospholipid antibodies, which recognize lipoidal material released from damaged host cells (cardiolipin, cholesterol and lecithin amongst others) and possibly from the treponemes (cardiolipin). Treponemal antibodies persist in the body for many years after treatment, whereas anti-phospholipid antibodies can be used to monitor seroreversion in cured patients. Hence, detection of a phospholipid-specific immune response is required to discriminate current infections from past exposures. However, as anti-phospholipid antibodies are found in several other infectious and non-infectious conditions, serological testing for both types of antibodies is currently recommended to diagnose acute syphilis infections. Ideally, an antibody detection test for HIV and syphilis (co-)infections should combine detection of antibodies to HIV-1, HIV-2, Treponemal and non-Treponemal antigens. In the special case of pregnant women a positive treponemal test result is indicative for infection and treatment is recommended. However, detection of non-treponemal antibodies could improve specificity as during pregnancy antibody levels tend to be elevated.
Developing a point-of-care test requires a step-by-step optimization of a complexity of variables that depend on the properties of the specimen, target, matrix components and detection reagents employed in a repetitive process of adjustment, validation and fine-tuning. Combined, these variables determine the requirements for the composition and properties of the assay fluids and procedure. Steps includes the selection and development of antigens, functionalization of the capture and detection probe, the incorporation (as dried stabilized reagent) of the probes in the test device, selection and preparation of test strip components, the development of reagents and fluid used for the preparation and application probes and components, and the development of the assay procedure. The development of LFA for single analytes has been standardized and equipment for LFA manufacturing is available. We first used LFA to select capture and detection reagents and to optimize conditions for the development of single analyte tests for HIV and syphilis antibody detection. Subsequently, the reagents and procedures for multiplexing and the development of the Cyclotek system were investigated. Reagents were first developed for visual (colloidal gold based) reading before testing for automatic read-out and use of fluorescent labels (Illustration 1).
Antigens selected for the development of capture and detection probes included HIV-1 p24, HIV-1 gp41, HIV-1 gp160 and HIV-2 gp39 recombinants for HIV testing, and the Treponema pallidum (syphilis) recombinant antigens TpN17 and TpN47 for syphilis testing. These HIV and Treponemal antigens exhibit major differences in biophysical, biochemical and antigenic properties that affected solubility and strongly influenced requirements both for binding to (solid) detection matrixes and conjugation to detection labels. Functionalized capture and (colloidal gold labelled) detection probes for HIV and syphilis testing revealed major differences in requirements for buffer system, pH and additives that limited the selection of a common reaction fluid with optimal test performance for combined detection. The optimal functioning of the HIV and Treponemal probes in LFA required assay fluids that differed in buffer system, pH and the presence of additives at values and concentrations that did not match, and thus attempts to combine the two systems for simultaneous HIV and syphilis detection adversely affected performance. Successive HIV/Syphilis LFA prototypes were developed to improve diagnostic performance. The HIV/syphilis LFA prototype 4 showed a sensitivity of 98.8% for HIV-1, 100% for HIV-2 and 72.7% for syphilis at a specificity of 97.7%, 100% and 90.9%, respectively, and the sensitivity for VDRL test positive samples classified as lues latens/after treatment/long existing lues was 84.2% and 100% for active syphilis/recently treated. An independent laboratory-based evaluation is ongoing. If desirable additional fine-tuning will be performed and standard operation procedures will be adjusted before a further field evaluation is planned. Combining HIV and syphilis testing in the Cyclotek system or the use of fluorescent labels required even more specific conditions that reduced test performance. Automatic read-out did add to standardization and for automatic reading of the HIV/Syphilis LFA the use of the Fio Deki reader may be considered.
Project Results:
Development and Evaluation of a HIV/Syphilis Combination Lateral Flow Assay with Visual Read Out
Several recombinant antigen were functionalized to obtain pairs of capture and detection probes and different combinations for HIV-1, HIV-2 and syphilis were tested in LFA. For each probe pair, optimal assay conditions were investigated, that including immobilisation (“striping”) conditions of the antigen onto the nitrocellulose test strip, conjugation to 40nm colloidal gold particles, composition of the assay fluid, and choice and treatment of conjugate and sample pads. Sample and conjugate pads are treated with various buffers to suppress non-specific binding of serum molecules and detection probes, to improve stability and release of the dried detection reagent and to ensure homogeneous flow. Results were evaluated for stability of the gold conjugate, appearance (homogeneous stained distinct line) of the staining at the test line, signal strength at the test line for positive control samples and absence of reactivity with negative control samples. Several available antigens developed by the PreventIt partners for use in ELISA showed no or poor functionality as detection and/or capture probes in the LFA and Cyclotek system. These antigens gave poor binding to the test matrix and/or instability of the detection reagent and a significant project effort was needed to re-develop, source and test HIV and syphilis antigens. An example of optimized capture probe diluents, assay fluid conditions and selected HIV-1, HIV-2 and Treponemal capture and detection probe combinations is presented in Table 1a. Notably, whereas for optimal detection of HIV-1 antibodies an HIV-1 specific capture probe was used in combination with a non-specific (Protein A) detection probe, analyte-specific capture and detection probes were used for HIV-2 and Treponemal antibody detection. Note that the assay fluids required for the optimal detection of antibodies to HIV and Treponemal antigens showed major differences in buffer base, pH and the requirement for additives. The multiplexing, first of HIV-1 and Treponemal antigen in Prototype 1 and subsequently of HIV-1, HIV-2 and Treponemal antigens in Prototype 2, 3 and 4, required major adjustments in the reaction fluid and the pretreatment of sample and conjugate pads for optimal functioning. Some of the essential adjustments that were implemented in the development of the various prototypes are listed in Table 1b. The development of the prototype tests is detailed in the Labjournals and (for Prototypes 3 and 4) documented in SOPs and essential development steps are summerized in the PreventIt product development file (documents on file at KITBR).
HIV/Syphilis LFA prototypes Of the several prototype HIV/Syphilis LFAs that were developed, four were promising and taken further for evaluation. The first prototype (Prototype 1) subjected to evaluation contained one test line for the detection of HIV-1 specific antibodies and one test line for the detection of Treponemal antibodies in addition to a reagent control line. The second prototype (Prototype 2) contained in addition to the two test lines for the detection of HIV-1 and Treponemal specific antibodies and the control line included in Prototype 1, a third test line for the detection of HIV-2 specific antibodies. Prototype 3 (like prototype 2) contained three test lines consisting of respectively HIV-1, HIV-2 and Treponemal capture probes and differed from Prototype 2 in the presence of two conjugate pads and in the treatment of sample and conjugate pads. Prototype 4 differed from prototype 3 by omitting the conjugate pad containing gp36/TpN17 and adding a non-treated conjugate pad to improve the migration of the gold conjugate over the test strip and to give a more homogeneous staining of the test lines. Also, the assay fluid was adjusted.
Serum samples and reference testing Prototype HIV/Syphilis LFAs were evaluated on a collection of 168 HIV and or Syphilis positive serum samples selected by the Amsterdam Municipal Health Services (GGD). Information on HIV and syphilis reference test results were provided by the Laboratory for STDs of the GGD. For syphilis detection the Treponema Screen Immunofluorescent EIA from Diasorin performed on the automatic Liaison XL analyser was used. Syphilis EIA positive samples (≥1.0) were confirmed by lues immunoblot. HIV testing was performed for all samples, syphilis EIA for all but 10 samples and lues immunoblot confirmatory testing for 94 EIA positive samples. Eighty samples were HIV test positive, 94 samples were syphilis EIA test positive and of these 77 were lues blot test positive. Staging of lues positive samples was further done by the combined results of EIA, Luesblot and VDRL testing. The test results of the reference tests are summarized in Table 2. Based in the syphilis serology samples were classified in a. no syphilis (EIA <0.1)(N=74) b. Dubious serology/no syphilis (EIA>0.1 immunoblot negative)(N=17), c. Lues latens/lues after treatment (EIA>0.1 immunoblot positive, VDRL negative)(N=37), d. Early lues/recently treated/long existing lues (EIA>0.1 immunoblot positive, VDRL <1:8)(N=19), e. Active lues/recently treated lues (EIA>0.1 immunoblot positive, VDRL ≥1:8)(N=13), f. Unknown (incomplete serology data). Anonymized samples were provided and patient specific information was not available.
HIV/Syphilis LFA Development and Validation
Testing of the prototype HIV/Syphilis LFAs was performed at the laboratory of KIT Biomedical Research and results were read by two readers.
• Prototype 1 (HIV-1 and Treponemal antibody detection) LFA
Test procedure The LFA was performed by the application of 5µl serum to the sample pad immediately followed by the application of 130µl assay fluid. Tests were examined for the presence of staining at the test and control lines after 15-20 minutes and test results were considered valid when staining was observed at the control line. Tests were considered positive for HIV-1 if staining was observed at the HIV-1 test line and considered positive for syphilis if staining was observed at the Treponemal test line. Staining intensity at the test lines was rated in comparison with a scoring card with clearly visible staining rated 1+ to 4+ with very weak staining scored 0.5+ (doubtful) and considered negative. The rapid test was performed on all 168 HIV and or syphilis positive samples randomly selected for HIV and or syphilis positivity by the GGD of Amsterdam.
Test results Prototype 1 LFA gave a valid result (staining at the control line) for all samples. A positive HIV-1 result was obtained for 76 (95%) of the 80 HIV positive serum samples by reader 1 and for 78 (97.5%) samples by reader 2. Of the 88 negative control samples two samples were scored positive by Reader 1 and three samples were scored positive by reader 2. Reading of the test results at the Syphilis test line by the two readers was also consistent. Reader 1 scored 63 (81.8%) out of 77 syphilis positive control samples positive and Reader 2 scored 62 (80.5%) of the positive control samples positive. The Syphilis EIA reference test results rated from ≥1.0 to >70 and the samples from the positive control group that scored negative at the Treponemal test line in the LFA had EIA test results <7.0. Of the negative control samples three samples were scored positive by the two readers. The test results are summarized in Table 3.
All 168 samples gave a valid test result in the Prototype 1 LFA (e.g. showing staining at the control line). Examples of Prototype 1 LFAs run with different specimens are presented in Figure 1.
Specific information on the HIV type for the HIV positive control samples obtained from the GGD was not available; it is assumed that all HIV positive samples were HIV-1 positive as this is the predominant HIV type present in the Dutch population. Prototype 1 tested negative for ten HIV-2 positive samples obtained from the Gambia (data not shown).
Rapid test staining intensity The staining intensity at the HIV-1 test line was rated weak (1+) to moderately strong (2+) for the majority of the HIV positive control samples. An on average stronger staining intensity was observed at the Treponemal test line (Table 4).
• Prototype 2 (HIV-1, HIV-2 and Treponemal specific antibody detection) LFA
Background information To include detection of HIV-2 infections a third test line containing HIV-2 specific antigen was added to the test zone of Prototype 1 resulting in Prototype 2. Prototype 2 also differs from Prototype 1 in the composition and amount of detection reagent, which was adjusted to improve signal intensity at the HIV-1 test line.
Test procedure The assay procedure for Prototype 2 was the same as for Prototype 1 and the test was evaluated on the same serum panel of 168 HIV and or Syphilis positive samples provided by the GGD of Amsterdam.
Test results Examples of Prototype 2 test results are presented in figure 2. The HIV-1 test line gave a relatively high sensitivity (Table 5) and specificity with strong staining for HIV-1 positive control samples (Table 6). The HIV-2 test line tested positive in 23 (reader 1) and 25 (reader 2) HIV positive sample indicating cross-reactivity with HIV-1 and unfortunately showed a poor specificity due to cross-reactivity with syphilis positive samples. Compared with Prototype 1 the Treponemal test line showed a somewhat reduced sensitivity (76.6% by reader 1 and 72.7% by reader 2). The specificity was high (96.1%). The HIV-2 test line reacted positive in 10 out of 10 HIV-2 positive samples obtained from the Gambia albeit at a relatively low staining intensity (Table 7). The HIV-1 test line showed no reactivity for these samples but six samples tested positive at the Treponemal test line. The presence of specific Treponemal antibodies in these samples was confirmed by the Syphilis 3.0 rapid test from Standard Diagnostics (Alere). In Prototype 2 significant cross-reactivity was observed with a number of HIV negative control samples. Further studies learned that cross-reaction at the HIV-2 test line in Prototype 2 could be suppressed by a slight modification in the assay fluid but changes in the composition of the assay fluid had a strong negative effect on the sensitivity, specificity and signal intensity at the HIV-1 and Treponemal test lines. The conversion of Prototype 1 to Prototype 2 also included changes in concentration of detection reagent and composition of the assay fluid resulting in a less strong reactivity at the Treponemal test line and a reduced detection rate of syphilis positive samples with low syphilis EIA titres of <7.0 (on a scale of 1.0 - >70 units).
• Prototype 3 (HIV-1, HIV-2 and Treponemal specific antibody detection) LFA
Background information Prototype 2 showed non-specific reactivity at the HIV-2 test line and a limited sensitivity at the Treponemal test line (see above). These issues were tackled by using a different HIV-2 antigen, increasing the concentration of the Treponemal detection probe, and using two conjugate pads, one for the HIV probe and one for the Treponemal detection probes, each pre-treated using a different solution before application of the (dried) detection reagent. A Protein A based detection reagent was added to the sample pad containing the HIV probe.
Test procedure The assay procedure for Prototype 3 was the same as for Prototypes 1 & 2 and the test was evaluated using the same serum panel of 168 HIV and or syphilis positive samples provided by the GGD of Amsterdam.
Test results The HIV-1 test line of Prototype 3 gave a relatively high sensitivity and specificity (Table 8) with strong staining for HIV-1 positive control samples (Table 9). The HIV-2 test line tested positive in 17 (reader 1) and 19 (reader 2) HIV-1 positive sample indicating cross-reactivity with HIV-1 but did not react with any of the syphilis positive samples showing a high specificity. Compared with Prototype 2 the Treponemal test line showed an improved sensitivity (90.9% by reader 1 and 84.4% by reader 2). However, the specificity was decreased (75.3% reader 1 and 71.4% reader 2) compared to Prototype 2. The HIV-2 test line in Prototype 3 reacted positive in 10 out of 10 HIV-2 positive samples obtained from the Gambia (Table 9) with increased staining intensity compared to Prototype 2 (Table 10). The HIV-1 test line showed no reactivity for these samples but 4 samples tested positive at the Treponemal test line. The presence of specific Treponemal antibodies in these samples was confirmed by the Syphilis 3.0 rapid test from Standard Diagnostics (6 out 10 testing positive for syphilis). Another observation made for this prototype was that test lines did not stain homogenously.
• Prototype 4 (HIV-1, HIV-2 and Treponemal specific antibody detection)
Background To improve the specificity at the Treponemal test line, the separate detection reagents that were applied to a separate conjugate pad were in this prototype omitted. Also, in prototype 4 an extra non-treated conjugate pad was added to improve the migration of the gold conjugate over the test strip which should result in more homogeneous flow and homogeneous staining of the test lines.
Rapid test procedure The test procedure for Prototype 4 was the same as for the previous Prototypes. The test was evaluated on the same serum panel of 168 HIV and or Syphilis positive samples provided by the GGD of Amsterdam.
Rapid test results The HIV-1 test line gave a relatively high sensitivity (Table 11) and specificity with strong staining for HIV-1 positive control samples (Table 12). Representative examples are presented in figure 3. The HIV-2 test line tested positive in 20 (reader 1) and 24 (reader 2) HIV positive sample indicating cross-reactivity with HIV-1 but did not react with only syphilis positive samples which indicates a high specificity. The HIV-2 test line reacted positive in 10 out of 10 HIV-2 positive samples obtained from the Gambia with increased staining intensity compared to prototype 2 (Table 13). The HIV-1 test line showed no reactivity for these samples but 4 samples tested positive at the Treponemal test line. The presence of specific Treponemal antibodies in these samples was confirmed by the Syphilis 3.0 rapid test from Standard Diagnostics (6 out 10 were positive for Syphilis). Compared with Prototype 3 the Treponemal test line showed a decreased sensitivity (72.7% by reader 1 and 67.5% by reader 2) however the specificity was increased (90.9% reader 1 and 92.2% reader 2). The majority of the syphilis positive samples that did not react in at the Treponemal test line in Prototype 4 had a EIA value <8.0 and were immunoblot positive. All but three of these samples did not react in the VDRL and were classified as earlier syphilis after treatment or Lues latens. The three sample that did react in the VDRL had a low but significant VDRL titre of 1:1 to 1:4, a low EIA value of <8.0 and was classified as early Lues/recently treated/long existing syphilis (84.2% sensitivity). None of these LFA test negative samples were from patients diagnosed with active Lues/Lues latens (e.g. EIA ≥1.1 immunoblot positive and VDRL test positive ≥1:8)(100% sensitivity).
Discussion Prototype 4 has a high sensitivity and specificity for HIV-1 and HIV-2. The changes converting Prototype 3 to Prototype 4 improved the specificity of the Treponemal test line but reduced sensitivity for syphilis positive samples with an EIA >1.0. Table 14 and figure 4 summarize the results obtained for Prototypes 1, 2, 3 and 4. The majority of the syphilis positive samples that did not react at the Treponemal test line in Prototype 4 reacted weakly at a very low value of <8.0 in the EIA and were from patients classified as earlier syphilis after treatment or lues latens based on a negative VDRL test. Three rapid test negative syphilis sample was classified as Lues latens/after treatment/long existing lues and the sensitivity for this group of patients was 84.2%. All 13 samples from patients with active syphilis tested positive (100% sensitivity) at the Treponemal test line and a more extensive validation on a larger panel of samples will be required to determine the sensitivity for patients with active syphilis more accurately. A further fine tuning of reagents may provide a better balance between the specificity and sensitivity at the syphilis test line. As pointed out by Causer and coworkers existing HIV/Syphilis combination tests lack specificity for the detection of Treponemal antibodies in HIV positive individuals (Causer et al PLOS One 2014, 9, e91504).
It should be noted that in all tests run (Prototypes 1, 2, 3 & 4) the control line showed clear staining showing that the test results were valid demonstrating robustness of the system.
Automatic read-out of HIV/Syphilis LFA
Background Reading of LFA test by visual inspection is prone to subjectivity in particular if the signal strength is weak and or reading is done under poor illumination or by users requiring reading aids. Automatic reading tests lines in LFA may ensure standardization (Figure 5).
Method A reader was developed and used to scan the centre part of the LFA strip (Figure 5a) To assess the sensitivity of automatic reading HIV/Syphilis LFA prototype 3 test strips were subject to reading with the reading device, and the results were compared with visual inspection by two readers. To avoid interference from signals due to the shape of the reading area the reader was programmed to scan the centre part of the test strip.
Results Results for a panel of 20 samples are presented in Table 15. Discrepant results were obtained for several samples that gave a 1+ reading intensity for either visual or automatic reading.
Discussion Digital signal detection method allows objective signal quantification and has a strong correlation with the visual detection at a staining intensity ≥1+. Further software upgrading and standardization will be needed to allow fully automatic data processing and interpretation. Adjustment of the LFA cassette may be needed to flatten background levels currently affected by the presence of the notch dividing the reading window in a test and control zone and shadow arising at the left end of the test strip due to positioning of the light source. Removing of the notch may allow positioning of the test lines more centrally in the reading window away from the sample and conjugate pad in the current cassette and away from the shadow causing a back ground signal interfering with the HIV-1 test line.
HIV/Syphilis Combination LFA Assay using fluorescent labelled detection probes for automatic read out
Background The use of fluorescent labels instead of colloidal gold based nanoparticles could improve sensitivity and facilitate automatic reading. Colloidal gold nanoparticles are relatively unstable and the use of fluorescent labels could add to the robustness of the system. For fluorescent detection in LFA, Treponemal and HIV antigens were conjugated to a series of well-established commercial fluorescent labels including AlexaFluor-430, europium chelate reagents such as CM-EU and Dragon Green, and novel fluorescent polymers such as MEHPPV (Poly [2-methoxy-5-(2-ethylhexyloxy)-1, 4-phenyllenevinylene] nanoparticles developed by Dr. Mark Green at King’s College London) and BV605 (Brilliant Violet 605, Sirigen Inc. (Beckton, Dickinson & Company)). In initial investigations, fluorescent labels such as the MEHPPV, AlexaFluor-430 and Dragon Green gave either no or poor signal intensity, significant background activity and/or poor reproducibility when applied in LFA. More consistent results were obtained in LFA with CM-Europium (CM-EU) and Sirigen’s BV605 label conjugated to Treponemal specific antigen. Of these two labels, Sirigen’s BV605 showed relatively strong signals at the Treponemal test line when tested with syphilis positive serum samples (Figure 6, samples 1 to 10) and no reactivity with a set of control samples (Figure 6, samples 11 to 15). These results were obtained using the conditions summarized in Table 17a. Treponemal antigen labelled with CM-EU gave positive (1+ - 4+) results with all but one of the syphilis positive samples and weak positive (1+) results with the negative control samples (data not shown). Gp41 labelled with CM-EU for HIV-1 antibody detection showed inconsistent results with very strong signals for some HIV positive samples and no or very weak signals for others. Also a variety of staining artefacts including aspecific binding developed at the edges of the detection line for HIV negative samples was observed (data not shown) that was not solved by testing a range of conditions. Therefore the use of this label was not further investigated and further work concentrated on probes labelled with Sirigen’s BV605.
Method To circumvent some problems such as poor solubility of the detection probe when using the Sirigen’s BV605 in LFA, tests were performed in a “dipstick” format with the test strip placed in a reaction vessel containing a mixture of 2µl detection probe, 5µl serum and 200µl assay fluid. Assays were allowed to develop for 20min after which the results is documented using the CCM reader. Probes and assay conditions were initially developed for using Treponemal probes. To include HIV-1 and HIV-2 detection, the GP36/41 HIV-1/-2 fusion protein was included as second capture probe in a test line next to the Treponemal test line. This capture probes functions as capture probe for an HIV-1 antibodies and HIV-2 antibodies combined in a single test line. To minimize cross-reactivity at the Treponemal test line, the concentration and ratio of the TpN17/TpN47 capture proteins were adjusted and a different nitrocellulose detection strip was used (see Table 17b). The test was performed on 43 samples of which 22 were HIV1 and or HIV2 positive, 12 were positive for syphilis and 9 samples were negative to both syphilis and HIV1/2.
Results The sensitivity of the HIV/Syphilis test was 100% for HIV1/2 samples and 91% for syphilis positive samples. The specificity was 85% and 93%, respectively. Examples of HIV-1, HIV-2 and syphilis positive and negative test results are presented in Figure 7. Results of a direct comparison of LFA with fluorescent reading with visual detection in Prototype 2 & 3 LFAs using colloidal gold labelled detection probes is presented in Table 18.
Discussion Multiplexing for combined HIV-1/2 and syphilis antibody detection required major adjustments in the composition of the assay fluid compared with conditions established for each of the two systems separately which led to a decrease in signal intensity and specificity. Despite the relatively weak signal intensity the sensitivity was relatively high. The specificity in particular was low for HIV testing. The use of fluorescent detection seems to require more stringent conditions than the use of colloidal gold nanoparticles, limiting options for multiplexing, and optimized conditions negatively affected signal intensity and specificity. With the BV605 fluorescent-labelled probes some positive patient specimens with similar EIA titres and giving a similar staining intensity in LFA with gold-labelled probes gave strong differences in signal intensity. This suggests that whilst showing promise due to the brightness of the fluorophore, the biophysical characteristics of the BV605 fluorescent conjugates would require further optimisation to meet the specific requirements for this challenging HIV/syphilis LFA application.
Development of the HIV/Syphilis Combination Cyclotek Assay
Based on experience gained with the development of a multi-analyte LFA for HIV/Syphilis antigens and probes were selected for use in the development of the Cyclotek system. The Cyclotek systems differs from the immunochromatographic LFA in several aspects and one of these is the use of a solid support on the surface of which the capture probe is immobilized and the reaction takes place. In addition to different matrix types, gamma irradiation and pre-treatment of the matrix with corrosive agents and the type and composition of the binding fluid has been investigated to optimize binding of the capture probe. The use of a solid matrix also required the development and optimisation of a new assay buffer to dilute the sample, dissolve the detection probe and un the assay.. Antigens used as capture and detection probes were GP41 recombinant antigen, and TpN17 and TpN47 recombinant antigens at a ratio of 7:3. Both TpN17 and TpN47 as plain antigen and after conjugation to BSA. The BSA conjugates increased the appearance of the stained detection spots but also reduced the signal intensity and their use was not further investigated. Examples of buffers tested in the development of a HIV-1/Syphilis Cyclotek system are listed in Table 19. Buffer 1 (1xPBS pH 7.2 supplemented with 3% Tween-20, and 1.67% BSA) gave optimal functioning of the TpN-17 capture and detection probes and buffer 14 (1% Borax pH 7.0 supplemented with 1.5% Tween 20, 1% NaCl) gave promising results for the HIV-1 capture and detection probes. Buffers 1 and 14 differ significantly in composition and while buffer 1 gave poor results for HIV-1 detection, buffer 14 gave poor results for syphilis antibody detection. Based on the experience with lateral flow a combination of the two buffers (buffer system 4) was used to combine the two detection systems. The resulting buffer was tested at a wide pH range with a panel of serum samples used for test optimization. Combining the two buffers reduced signal intensity compared with buffers 1 and 14. The strongest signal intensity for the two analytes was obtained for buffer 4 adjusted to pH7.2 and buffer 4 with pH7.2 was therefore used for further test evaluation.
Test procedure The reaction in Prototype Cyclotek assay is started by the application of 175µl reaction mixture consisting of 150µl assay fluid (1xPBS, 0.5% borax, 2.25% Tween 20, 0.84% BSA, 0.5% NaCl pH7.2) 5µl detection reagent (GP160 4,5mg/l and TpN17 9mg/l conjugated to 40nm colloid gold, OD20) and 20µl sample onto the PE test surface of the Cyclotek. Tests were rotated for 15 min at a rotation speed of 5rpm. After 15 min the reaction mixture was removed and replaced with 150µl assay fluid and the rotation of the test disk was continued for 1 min. After removal of the test fluid the tests were examined by visual inspection for the presence of staining at the test dots (Figure 8). Tests were considered positive for HIV-1 when staining was observed at the HIV-1 test line and considered positive for syphilis when staining was observed at the Treponemal test line. Staining intensity at the test lines was rated in comparison with a scoring card with clearly visible staining scored 1+ to 4+ and considered positive, and very weak staining rated 0.5+ (doubtful) and considered negative. The rapid test was performed on 52 HIV positive and 49 syphilis positive samples obtained from BBI Solutions (Cardiff, UK). Information on the syphilis status was not available for the HIV positive serum samples and information on the HIV status was not available for the syphilis positive samples. After reading of the test result by visual inspection, disks were read digitally in the CCM reader.
Test results The main limitations of the optimized Cyclotek system are the low signal intensity obtained after 75 rotation rounds (15 min reaction time) and the presence of non-specific activity (Table 20). The signal intensity at the HIV detection spot was in general lower than the signal intensity at the Treponemal test spot and resulted in a relatively poor sensitivity for HIV (84.5%) compared with 100% for Treponemal antibody detection. Possibly the degree of false positive results (85.7% and 86.5% specificity, respectively) is overestimated as the HIV status of the syphilis positive control samples and the syphilis status of the HIV positive control samples was not available.
Conclusion Further options for optimization of the test conditions are limited due to the specific buffer requirements of two analyte systems but might be found in the use of another solid support and or the use of alternative detection reagents and blocking reagents.
HIV/Syphilis Combination Cyclotek Assay with automatic read out
A reader was constructed consisting of an AVT Mako G-125C color camera in combination with a 25mm Techspec Compact fixed focal lens. Dependent on the selected settings, the sample is illuminated with violet or white light (Figure 10). A simple automatic detection algorithm consisting of an image pre-processing and spot detection block is used to detect spots. For red spot detection created by colloidal gold labelled detection probes under white light illumination the difference between the red and green signal intensity is enhanced resulting in a better observable spot. Results of automatic reading was compared with visual reading of wet spots observed directly after the end of the incubation and washing procedure (1) and with visual reading after the remaining liquid left after the final washing step of the test procedure was dried (2). Automatic reading was performed on the dried disk with optimal reading settings for HIV and syphilis detection (3). However, even with these optimal settings, visual read out by method 1 and 2 result in a better accuracy than method 3 (Figure 11).
Discussion and Summary
A multi-analyte LFA was developed for the serodiagnosis of HIV and syphilis. The HIV1,2/syphilis LFA prototype 4 had a sensitivity of 98.8% for HIV-1, 100% for HIV-2 and 72.7% for syphilis at a specificity of 97.7%, 100% and 90.9%, respectively, and the sensitivity for VDRL test positive samples classified as lues latens/after treatment/long existing lues was 84.2% and 100% for active syphilis/recently treated. The diagnostic accuracy of the combination assay for HIV was similar to that of existing HIV/Syphilis POC tests.1,2 An asset of our HIV1,2/Syphilis LFA is that it allows discrimination between HIV-1 and HIV-2 infections. The diagnostic accuracy of the HIV1,2/Syphilis LFA for the serodiagnosis of syphilis may need further fine-tuning as in different prototypes the sensitivity and specificity varied and were somewhat lower than reported for existing LFAs. In published studies the Treponemal Pallidum Particle Agglutination test was used as comparator and whereas we used a highly sensitive EIA3 as reference, this could explain the observed lower sensitivity for our LFA. The majority of the LFA negative test results obtained for EIA positive samples had relatively low EIA values <7.0 that tested lues blot positive but VDRL test negative and that were classified as earlier infection after treatment/lues latens if never treated. A further independent laboratory based evaluation study is ongoing to confirm the diagnostic performance of the HIV1,2/Syphilis LFA prototype 4. The standard operation procedures for the HIV1,2/Syphilis LFA Prototype 4 have been documented and a manufacturing file is available.
Fluorescent labels were tested to determine whether the use of fluorescent detection probes could improve test performance. Of the different fluorescent probes tested Sirigen’s BV605 gave the most promising results in LFA and a HIV/Syphilis dipstick was developed with a sensitivity of 100% for HIV1/2 samples and of 91% for syphilis positive samples. The specificity was 85% and 93%, respectively. In this dipstick test liquid detection probe was used, and for conversion to LFA it would be necessary to establish conditions for drying and conservation of the fluorescent labelled probes onto the conjugate pad of the LFA device and that allow rapid release of functional detection reagent by the application of the sample and assay fluid mixture. Whereas Sirigen’s BV605 provided a strong and specific signal the inability to identify a common assay fluid for the optimal serodiagnosis of HIV and syphilis in one assay limited its use in the development of a HIV/syphilis combination assay.
Options for optimization of the HIV1,2/Syphilis LFA test performance were restricted by the differences in requirements in buffer and buffer additives for the optimal functioning of the two detection systems. Requirements for the optimal performance of the two detection systems in the Cyclotek system appeared to be even more sensitive to the choice of assay components and reagents and the composition of the assay fluid thus further restricted options for optimization. A major factor that hindered test development was the poor binding of the capture probes to the solid support. Whereas relatively strong signals were obtained for syphilis positive control samples, weaker signals were obtained for HIV positive control samples and the signal strength at the HIV capture probe would need to be improved to improve specificity. Automatic read-out of Cyclotek test signals did contribute to standardized test interpretation but did not improve sensitivity.
The use of existing readers developed for LFA such as the Fio Deki reader may facilitate standardized reading of the HIV/Syphilis LFA developed here.5,6 The use of a reader may also improve sensitivity as is observed for the reading of malaria POC testing.
Literature references
1. Bristow CC, Adu-Sarkodie Y, Ondondo RO et al. Multisite Laboratory Evaluation of a Dual Human Immunodeficiency Virus (HIV)/Syphilis Point-of-Care Rapid Test for Simultaneous Detection of HIV and Syphilis Infection. Open Forum Infect Dis. 2014;1:ofu015.
2. Yin YP, Ngige E, Anyaike C et al. Laboratory evaluation of three dual rapid diagnostic tests for HIV and syphilis in China and Nigeria. Int J Gynaecol Obstet. 2015;Suppl 1:S22-6.
3. Knight CS1, Crum MA, Hardy RW. Evaluation of the LIAISON chemiluminescence immunoassay for diagnosis of syphilis. Clin Vaccine Immunol. 2007;14:710-3.
4. Causer LM, Kaldor JM, Fairley CK et al. A laboratory-based evaluation of four rapid point-of-care tests for syphilis. PLoS One. 2014 Mar 11;9(3):e91504.
5. Shekalaghe S, Cancino M, Mavere C, Juma O, Mohammed A, Abdulla S, Ferro S. Clinical performance of an automated reader in interpreting malaria rapid diagnostic tests in Tanzania. Malar J. 2013;12:141.
6. Herrera S, Vallejo AF, Quintero JP et al. Field evaluation of an automated RDT reader and data management device for Plasmodium falciparum/Plasmodium vivax malaria in endemic areas of Colombia. Malar J. 2014;13:87.
Potential Impact:
Due to limited awareness among pregnant women and health care providers, lack of availability of simple and rapid diagnostics tests that combine HIV and syphilis detection and lack of access to diagnostic facilities where diagnostic testing for HIV and syphilis may be performed many pregnant women at risk of infection with either or both of the pathogens do not receive an appropriate diagnosis and are not treated. As a result, the risk of congenital transmission with severe damage to the child remains high. The wider availability of affordable high quality point-of-care diagnostics will contribute to improved awareness and a better diagnosis and treatment. The development of the HIV1,2/Syphilis LFA developed by the PreventIt team is the first step towards a better access of pregnant women to diagnosis and treatment for HIV and syphilis. Unfortunately, due to a management decision to stop all laboratory activities of partner 2 the diagnostic performance of the HIV1,2/Syphilis LFA could not be confirmed in a field study by testing pregnant women at risk of HIV and Syphilis infection attending a mother and child clinic in sub-saharan Africa. As a result, essential data is lacking to bring the test to the market. Some of the laboratory activities of partner 1 on point of care test development will be taken over by the newly established Mondial Diagnostics Foundation. This foundation was established to develop test and to bring test more efficiently to the market and to serve patients more effectively.
Mother to child transmission of syphilis and HIV are major causes of severe morbidity and early death of the young infant. Treatment of the mother reduces the risk of congenital transmission. Co-infections are common and most cases of congenital transmission of HIV and syphilis infection take place in hard-to-reach populations, where the availability of an accurate diagnostic test for onsite screening of pregnant women is essential to gain access to and to deliver tailored medical care. The use of a point-of-care test for the combined detection of HIV and syphilis infections in pregnant women would allow same day testing and treatment thus reducing the number of patients that are lost during follow-up and increasing the number of infected persons that are receiving treatment.
In this project the application of two different test platforms for point-of-care testing for HIV and syphilis infections was investigated. The first platform is the lateral flow assay (LFA) while the second test format (Cyclotek) utilizes a platform specifically designed for multi-analyte testing. In this platform, the complex assay strip that forms the heart of the LFA and is composed of a number of overlapping components is replaced by a single moulded solid matrix with a circular groove on which the reaction takes place. The use of a single component matrix simplifies manufacturing and whereas space in LFA for the application of capture probes is limited, multiple capture probes can be immobilized at defined positions in the Cyclotek reaction groove. Moreover, potential for improved sensitivity of the Cyclotek assay compared to LFA was anticipated; the circular design permitting multiple rounds of interaction between serum molecules and probes. Using these two platforms, the multiplexing of HIV-1, HIV-2, Treponemal and non-Treponemal probes for the detection of HIV and syphilis antibody responses were investigated. To improve test performance and standardisation through operator independent automated readout, the use of fluorescent probes was investigated in addition to conventional visual detection.
Options for multiplexing HIV and syphilis capture and colloidal gold-labelled detection probes appeared to be restricted due to the different biophysical properties of the syphilis and highly hydrophobic HIV antigens and the resulting narrow range in assay conditions for the optimal functioning of individual probe combinations. In particular assay conditions required for non-Treponemal (cardiolipin) antibody testing turned out to be difficult to combine effectively. Combining HIV-1 and HIV-2 antibody testing with Treponemal specific antibody in LFA with visual read-out gave a sensitivity of 98.8% for HIV-1, 100% for HIV-2 and 72.7% for syphilis at a specificity of 97.7%, 100% and 90.9%, respectively. The majority of the syphilis positive samples that did not react had low values in the reference EIA and were classified as earlier syphilis after treatment or lues latens based on a negative Venereal Disease Research Laboratory (VDRL) test. The sensitivity for VDRL test positive samples classified as lues latens/after treatment/long existing lues was 84.2% and 100% for active syphilis/recently treated. An independent evaluation to confirm results is being performed and further fine-tuning may improve the diagnostic accuracy of the syphilis components. The development of the Cyclotek system appeared to be even more challenging. Conditions for HIV testing did not match those for syphilis and conditions for combined detection resulted in a reduced signal intensity for HIV with cross-reactivity observed between HIV and syphilis probes resulting in a suboptimal diagnostic performance. The use of fluorescent detection probes did not improve test performance. Of various fluorescent labels tested, including novel fluorescent polymers and a number of well-established commercial benchmark fluorophores Sirigen’s BV605 gave the brightest signals and is a very promising detection label for LFA. However, due to the characteristics of the HIV and Treponemal antigens, the conditions required for combined detection reduced the signal strength of the fluorophore, in turn reducing assay sensitivity and specificity. The use of automated reading adds to standardization, in particular for specimens that give a low or borderline signal intensity, but did not improve diagnostic accuracy.
Standard operation procedures have been prepared for the HIV/Syphilis LFA prototype assay and product manufacturing file highlighting essential steps in the development of the assay is available.
For field evaluation in pregnant women in mother and child clinics in sub-Saharan Africa an evaluation batch was prepared and documents to obtain ethical approval were prepared. Due to an unfortunate management decision of partner 1 the project was ended before the anticipated end date and field studies to confirm the excellent diagnostic performance in the target group and to develop guidelines for best practices could not be performed.
Project Context and Objectives:
Current point-of-care diagnostics for infectious diseases are based mainly on the immunochromatographic lateral flow assay (LFA) format which technology. The number of probes for different pathogens that can be combined in a single LFA is limited by the design of the assay and the physical format of the LFA. Also, in LFA the clinical specimen and detection reagent migrate over the test line once and are allowed to react with the immobilized capture probe during a short period, potentially limiting sensitivity. The Cyclotek platform is specifically designed for multi-analyte detection by using a single solid detection surface that allows the deposition of capture probes at multiple positions. The Cyclotek system may also improve sensitivity as the system is designed for multiple rounds of interactions between the capture probes, serum and detection reagent.
LFA point-of-care tests that utilize coloured detection reagents such as colloidal gold nanoparticles or stained latex particles enable reading of test results by visual inspection. The contrast of the detection reagent limits sensitivity and visual inspection of LFA test results affects reproducibility, especially at the lower end of the detection limit. Reading of test results by visual inspection is greatly influenced by (artificial) lighting conditions and found more difficult by users wearing reading glasses – thus the current trend directs towards the use of readers for standardized LFA test reading and interpretation. The use of fluorescent labels may improve sensitivity and simplify automatic read-out. If linked to an online database the use of readers provides many other advantages that benefits the health system and patient care such as improved patient registration and disease surveillance.
The PreventIt project focuses on the development of a point-of-care test for the serodiagnosis of HIV and syphilis infections as a model for a multi-analyte testing. HIV and syphilis infections during pregnancy pose high risks for the infant. Their timely treatment in pregnant women prevents congenital transmission and reduces the risk of an adverse pregnancy outcome. Co-infection are also very common among other risk groups. Notwithstanding, pregnant women with a high risk of infection do not always receive the required medical attention. Often, if testing is available it is done for only one of the two infections. Thus, the use of point-of-care testing is essential to provide diagnostic services to such hard-to-reach groups, allowing same day testing and thereby reducing the number of infected individuals who do not receive treatment.
A test for HIV infections should include reagents for the detection of antibodies against different HIV types, in particular when used in the African context. The immune response in syphilis involves production of Treponemal antibodies, which are specific for T. pallidum antigens, and non-Treponemal anti-phospholipid antibodies, which recognize lipoidal material released from damaged host cells (cardiolipin, cholesterol and lecithin amongst others) and possibly from the treponemes (cardiolipin). Treponemal antibodies persist in the body for many years after treatment, whereas anti-phospholipid antibodies can be used to monitor seroreversion in cured patients. Hence, detection of a phospholipid-specific immune response is required to discriminate current infections from past exposures. However, as anti-phospholipid antibodies are found in several other infectious and non-infectious conditions, serological testing for both types of antibodies is currently recommended to diagnose acute syphilis infections. Ideally, an antibody detection test for HIV and syphilis (co-)infections should combine detection of antibodies to HIV-1, HIV-2, Treponemal and non-Treponemal antigens. In the special case of pregnant women a positive treponemal test result is indicative for infection and treatment is recommended. However, detection of non-treponemal antibodies could improve specificity as during pregnancy antibody levels tend to be elevated.
Developing a point-of-care test requires a step-by-step optimization of a complexity of variables that depend on the properties of the specimen, target, matrix components and detection reagents employed in a repetitive process of adjustment, validation and fine-tuning. Combined, these variables determine the requirements for the composition and properties of the assay fluids and procedure. Steps includes the selection and development of antigens, functionalization of the capture and detection probe, the incorporation (as dried stabilized reagent) of the probes in the test device, selection and preparation of test strip components, the development of reagents and fluid used for the preparation and application probes and components, and the development of the assay procedure. The development of LFA for single analytes has been standardized and equipment for LFA manufacturing is available. We first used LFA to select capture and detection reagents and to optimize conditions for the development of single analyte tests for HIV and syphilis antibody detection. Subsequently, the reagents and procedures for multiplexing and the development of the Cyclotek system were investigated. Reagents were first developed for visual (colloidal gold based) reading before testing for automatic read-out and use of fluorescent labels (Illustration 1).
Antigens selected for the development of capture and detection probes included HIV-1 p24, HIV-1 gp41, HIV-1 gp160 and HIV-2 gp39 recombinants for HIV testing, and the Treponema pallidum (syphilis) recombinant antigens TpN17 and TpN47 for syphilis testing. These HIV and Treponemal antigens exhibit major differences in biophysical, biochemical and antigenic properties that affected solubility and strongly influenced requirements both for binding to (solid) detection matrixes and conjugation to detection labels. Functionalized capture and (colloidal gold labelled) detection probes for HIV and syphilis testing revealed major differences in requirements for buffer system, pH and additives that limited the selection of a common reaction fluid with optimal test performance for combined detection. The optimal functioning of the HIV and Treponemal probes in LFA required assay fluids that differed in buffer system, pH and the presence of additives at values and concentrations that did not match, and thus attempts to combine the two systems for simultaneous HIV and syphilis detection adversely affected performance. Successive HIV/Syphilis LFA prototypes were developed to improve diagnostic performance. The HIV/syphilis LFA prototype 4 showed a sensitivity of 98.8% for HIV-1, 100% for HIV-2 and 72.7% for syphilis at a specificity of 97.7%, 100% and 90.9%, respectively, and the sensitivity for VDRL test positive samples classified as lues latens/after treatment/long existing lues was 84.2% and 100% for active syphilis/recently treated. An independent laboratory-based evaluation is ongoing. If desirable additional fine-tuning will be performed and standard operation procedures will be adjusted before a further field evaluation is planned. Combining HIV and syphilis testing in the Cyclotek system or the use of fluorescent labels required even more specific conditions that reduced test performance. Automatic read-out did add to standardization and for automatic reading of the HIV/Syphilis LFA the use of the Fio Deki reader may be considered.
Project Results:
Development and Evaluation of a HIV/Syphilis Combination Lateral Flow Assay with Visual Read Out
Several recombinant antigen were functionalized to obtain pairs of capture and detection probes and different combinations for HIV-1, HIV-2 and syphilis were tested in LFA. For each probe pair, optimal assay conditions were investigated, that including immobilisation (“striping”) conditions of the antigen onto the nitrocellulose test strip, conjugation to 40nm colloidal gold particles, composition of the assay fluid, and choice and treatment of conjugate and sample pads. Sample and conjugate pads are treated with various buffers to suppress non-specific binding of serum molecules and detection probes, to improve stability and release of the dried detection reagent and to ensure homogeneous flow. Results were evaluated for stability of the gold conjugate, appearance (homogeneous stained distinct line) of the staining at the test line, signal strength at the test line for positive control samples and absence of reactivity with negative control samples. Several available antigens developed by the PreventIt partners for use in ELISA showed no or poor functionality as detection and/or capture probes in the LFA and Cyclotek system. These antigens gave poor binding to the test matrix and/or instability of the detection reagent and a significant project effort was needed to re-develop, source and test HIV and syphilis antigens. An example of optimized capture probe diluents, assay fluid conditions and selected HIV-1, HIV-2 and Treponemal capture and detection probe combinations is presented in Table 1a. Notably, whereas for optimal detection of HIV-1 antibodies an HIV-1 specific capture probe was used in combination with a non-specific (Protein A) detection probe, analyte-specific capture and detection probes were used for HIV-2 and Treponemal antibody detection. Note that the assay fluids required for the optimal detection of antibodies to HIV and Treponemal antigens showed major differences in buffer base, pH and the requirement for additives. The multiplexing, first of HIV-1 and Treponemal antigen in Prototype 1 and subsequently of HIV-1, HIV-2 and Treponemal antigens in Prototype 2, 3 and 4, required major adjustments in the reaction fluid and the pretreatment of sample and conjugate pads for optimal functioning. Some of the essential adjustments that were implemented in the development of the various prototypes are listed in Table 1b. The development of the prototype tests is detailed in the Labjournals and (for Prototypes 3 and 4) documented in SOPs and essential development steps are summerized in the PreventIt product development file (documents on file at KITBR).
HIV/Syphilis LFA prototypes Of the several prototype HIV/Syphilis LFAs that were developed, four were promising and taken further for evaluation. The first prototype (Prototype 1) subjected to evaluation contained one test line for the detection of HIV-1 specific antibodies and one test line for the detection of Treponemal antibodies in addition to a reagent control line. The second prototype (Prototype 2) contained in addition to the two test lines for the detection of HIV-1 and Treponemal specific antibodies and the control line included in Prototype 1, a third test line for the detection of HIV-2 specific antibodies. Prototype 3 (like prototype 2) contained three test lines consisting of respectively HIV-1, HIV-2 and Treponemal capture probes and differed from Prototype 2 in the presence of two conjugate pads and in the treatment of sample and conjugate pads. Prototype 4 differed from prototype 3 by omitting the conjugate pad containing gp36/TpN17 and adding a non-treated conjugate pad to improve the migration of the gold conjugate over the test strip and to give a more homogeneous staining of the test lines. Also, the assay fluid was adjusted.
Serum samples and reference testing Prototype HIV/Syphilis LFAs were evaluated on a collection of 168 HIV and or Syphilis positive serum samples selected by the Amsterdam Municipal Health Services (GGD). Information on HIV and syphilis reference test results were provided by the Laboratory for STDs of the GGD. For syphilis detection the Treponema Screen Immunofluorescent EIA from Diasorin performed on the automatic Liaison XL analyser was used. Syphilis EIA positive samples (≥1.0) were confirmed by lues immunoblot. HIV testing was performed for all samples, syphilis EIA for all but 10 samples and lues immunoblot confirmatory testing for 94 EIA positive samples. Eighty samples were HIV test positive, 94 samples were syphilis EIA test positive and of these 77 were lues blot test positive. Staging of lues positive samples was further done by the combined results of EIA, Luesblot and VDRL testing. The test results of the reference tests are summarized in Table 2. Based in the syphilis serology samples were classified in a. no syphilis (EIA <0.1)(N=74) b. Dubious serology/no syphilis (EIA>0.1 immunoblot negative)(N=17), c. Lues latens/lues after treatment (EIA>0.1 immunoblot positive, VDRL negative)(N=37), d. Early lues/recently treated/long existing lues (EIA>0.1 immunoblot positive, VDRL <1:8)(N=19), e. Active lues/recently treated lues (EIA>0.1 immunoblot positive, VDRL ≥1:8)(N=13), f. Unknown (incomplete serology data). Anonymized samples were provided and patient specific information was not available.
HIV/Syphilis LFA Development and Validation
Testing of the prototype HIV/Syphilis LFAs was performed at the laboratory of KIT Biomedical Research and results were read by two readers.
• Prototype 1 (HIV-1 and Treponemal antibody detection) LFA
Test procedure The LFA was performed by the application of 5µl serum to the sample pad immediately followed by the application of 130µl assay fluid. Tests were examined for the presence of staining at the test and control lines after 15-20 minutes and test results were considered valid when staining was observed at the control line. Tests were considered positive for HIV-1 if staining was observed at the HIV-1 test line and considered positive for syphilis if staining was observed at the Treponemal test line. Staining intensity at the test lines was rated in comparison with a scoring card with clearly visible staining rated 1+ to 4+ with very weak staining scored 0.5+ (doubtful) and considered negative. The rapid test was performed on all 168 HIV and or syphilis positive samples randomly selected for HIV and or syphilis positivity by the GGD of Amsterdam.
Test results Prototype 1 LFA gave a valid result (staining at the control line) for all samples. A positive HIV-1 result was obtained for 76 (95%) of the 80 HIV positive serum samples by reader 1 and for 78 (97.5%) samples by reader 2. Of the 88 negative control samples two samples were scored positive by Reader 1 and three samples were scored positive by reader 2. Reading of the test results at the Syphilis test line by the two readers was also consistent. Reader 1 scored 63 (81.8%) out of 77 syphilis positive control samples positive and Reader 2 scored 62 (80.5%) of the positive control samples positive. The Syphilis EIA reference test results rated from ≥1.0 to >70 and the samples from the positive control group that scored negative at the Treponemal test line in the LFA had EIA test results <7.0. Of the negative control samples three samples were scored positive by the two readers. The test results are summarized in Table 3.
All 168 samples gave a valid test result in the Prototype 1 LFA (e.g. showing staining at the control line). Examples of Prototype 1 LFAs run with different specimens are presented in Figure 1.
Specific information on the HIV type for the HIV positive control samples obtained from the GGD was not available; it is assumed that all HIV positive samples were HIV-1 positive as this is the predominant HIV type present in the Dutch population. Prototype 1 tested negative for ten HIV-2 positive samples obtained from the Gambia (data not shown).
Rapid test staining intensity The staining intensity at the HIV-1 test line was rated weak (1+) to moderately strong (2+) for the majority of the HIV positive control samples. An on average stronger staining intensity was observed at the Treponemal test line (Table 4).
• Prototype 2 (HIV-1, HIV-2 and Treponemal specific antibody detection) LFA
Background information To include detection of HIV-2 infections a third test line containing HIV-2 specific antigen was added to the test zone of Prototype 1 resulting in Prototype 2. Prototype 2 also differs from Prototype 1 in the composition and amount of detection reagent, which was adjusted to improve signal intensity at the HIV-1 test line.
Test procedure The assay procedure for Prototype 2 was the same as for Prototype 1 and the test was evaluated on the same serum panel of 168 HIV and or Syphilis positive samples provided by the GGD of Amsterdam.
Test results Examples of Prototype 2 test results are presented in figure 2. The HIV-1 test line gave a relatively high sensitivity (Table 5) and specificity with strong staining for HIV-1 positive control samples (Table 6). The HIV-2 test line tested positive in 23 (reader 1) and 25 (reader 2) HIV positive sample indicating cross-reactivity with HIV-1 and unfortunately showed a poor specificity due to cross-reactivity with syphilis positive samples. Compared with Prototype 1 the Treponemal test line showed a somewhat reduced sensitivity (76.6% by reader 1 and 72.7% by reader 2). The specificity was high (96.1%). The HIV-2 test line reacted positive in 10 out of 10 HIV-2 positive samples obtained from the Gambia albeit at a relatively low staining intensity (Table 7). The HIV-1 test line showed no reactivity for these samples but six samples tested positive at the Treponemal test line. The presence of specific Treponemal antibodies in these samples was confirmed by the Syphilis 3.0 rapid test from Standard Diagnostics (Alere). In Prototype 2 significant cross-reactivity was observed with a number of HIV negative control samples. Further studies learned that cross-reaction at the HIV-2 test line in Prototype 2 could be suppressed by a slight modification in the assay fluid but changes in the composition of the assay fluid had a strong negative effect on the sensitivity, specificity and signal intensity at the HIV-1 and Treponemal test lines. The conversion of Prototype 1 to Prototype 2 also included changes in concentration of detection reagent and composition of the assay fluid resulting in a less strong reactivity at the Treponemal test line and a reduced detection rate of syphilis positive samples with low syphilis EIA titres of <7.0 (on a scale of 1.0 - >70 units).
• Prototype 3 (HIV-1, HIV-2 and Treponemal specific antibody detection) LFA
Background information Prototype 2 showed non-specific reactivity at the HIV-2 test line and a limited sensitivity at the Treponemal test line (see above). These issues were tackled by using a different HIV-2 antigen, increasing the concentration of the Treponemal detection probe, and using two conjugate pads, one for the HIV probe and one for the Treponemal detection probes, each pre-treated using a different solution before application of the (dried) detection reagent. A Protein A based detection reagent was added to the sample pad containing the HIV probe.
Test procedure The assay procedure for Prototype 3 was the same as for Prototypes 1 & 2 and the test was evaluated using the same serum panel of 168 HIV and or syphilis positive samples provided by the GGD of Amsterdam.
Test results The HIV-1 test line of Prototype 3 gave a relatively high sensitivity and specificity (Table 8) with strong staining for HIV-1 positive control samples (Table 9). The HIV-2 test line tested positive in 17 (reader 1) and 19 (reader 2) HIV-1 positive sample indicating cross-reactivity with HIV-1 but did not react with any of the syphilis positive samples showing a high specificity. Compared with Prototype 2 the Treponemal test line showed an improved sensitivity (90.9% by reader 1 and 84.4% by reader 2). However, the specificity was decreased (75.3% reader 1 and 71.4% reader 2) compared to Prototype 2. The HIV-2 test line in Prototype 3 reacted positive in 10 out of 10 HIV-2 positive samples obtained from the Gambia (Table 9) with increased staining intensity compared to Prototype 2 (Table 10). The HIV-1 test line showed no reactivity for these samples but 4 samples tested positive at the Treponemal test line. The presence of specific Treponemal antibodies in these samples was confirmed by the Syphilis 3.0 rapid test from Standard Diagnostics (6 out 10 testing positive for syphilis). Another observation made for this prototype was that test lines did not stain homogenously.
• Prototype 4 (HIV-1, HIV-2 and Treponemal specific antibody detection)
Background To improve the specificity at the Treponemal test line, the separate detection reagents that were applied to a separate conjugate pad were in this prototype omitted. Also, in prototype 4 an extra non-treated conjugate pad was added to improve the migration of the gold conjugate over the test strip which should result in more homogeneous flow and homogeneous staining of the test lines.
Rapid test procedure The test procedure for Prototype 4 was the same as for the previous Prototypes. The test was evaluated on the same serum panel of 168 HIV and or Syphilis positive samples provided by the GGD of Amsterdam.
Rapid test results The HIV-1 test line gave a relatively high sensitivity (Table 11) and specificity with strong staining for HIV-1 positive control samples (Table 12). Representative examples are presented in figure 3. The HIV-2 test line tested positive in 20 (reader 1) and 24 (reader 2) HIV positive sample indicating cross-reactivity with HIV-1 but did not react with only syphilis positive samples which indicates a high specificity. The HIV-2 test line reacted positive in 10 out of 10 HIV-2 positive samples obtained from the Gambia with increased staining intensity compared to prototype 2 (Table 13). The HIV-1 test line showed no reactivity for these samples but 4 samples tested positive at the Treponemal test line. The presence of specific Treponemal antibodies in these samples was confirmed by the Syphilis 3.0 rapid test from Standard Diagnostics (6 out 10 were positive for Syphilis). Compared with Prototype 3 the Treponemal test line showed a decreased sensitivity (72.7% by reader 1 and 67.5% by reader 2) however the specificity was increased (90.9% reader 1 and 92.2% reader 2). The majority of the syphilis positive samples that did not react in at the Treponemal test line in Prototype 4 had a EIA value <8.0 and were immunoblot positive. All but three of these samples did not react in the VDRL and were classified as earlier syphilis after treatment or Lues latens. The three sample that did react in the VDRL had a low but significant VDRL titre of 1:1 to 1:4, a low EIA value of <8.0 and was classified as early Lues/recently treated/long existing syphilis (84.2% sensitivity). None of these LFA test negative samples were from patients diagnosed with active Lues/Lues latens (e.g. EIA ≥1.1 immunoblot positive and VDRL test positive ≥1:8)(100% sensitivity).
Discussion Prototype 4 has a high sensitivity and specificity for HIV-1 and HIV-2. The changes converting Prototype 3 to Prototype 4 improved the specificity of the Treponemal test line but reduced sensitivity for syphilis positive samples with an EIA >1.0. Table 14 and figure 4 summarize the results obtained for Prototypes 1, 2, 3 and 4. The majority of the syphilis positive samples that did not react at the Treponemal test line in Prototype 4 reacted weakly at a very low value of <8.0 in the EIA and were from patients classified as earlier syphilis after treatment or lues latens based on a negative VDRL test. Three rapid test negative syphilis sample was classified as Lues latens/after treatment/long existing lues and the sensitivity for this group of patients was 84.2%. All 13 samples from patients with active syphilis tested positive (100% sensitivity) at the Treponemal test line and a more extensive validation on a larger panel of samples will be required to determine the sensitivity for patients with active syphilis more accurately. A further fine tuning of reagents may provide a better balance between the specificity and sensitivity at the syphilis test line. As pointed out by Causer and coworkers existing HIV/Syphilis combination tests lack specificity for the detection of Treponemal antibodies in HIV positive individuals (Causer et al PLOS One 2014, 9, e91504).
It should be noted that in all tests run (Prototypes 1, 2, 3 & 4) the control line showed clear staining showing that the test results were valid demonstrating robustness of the system.
Automatic read-out of HIV/Syphilis LFA
Background Reading of LFA test by visual inspection is prone to subjectivity in particular if the signal strength is weak and or reading is done under poor illumination or by users requiring reading aids. Automatic reading tests lines in LFA may ensure standardization (Figure 5).
Method A reader was developed and used to scan the centre part of the LFA strip (Figure 5a) To assess the sensitivity of automatic reading HIV/Syphilis LFA prototype 3 test strips were subject to reading with the reading device, and the results were compared with visual inspection by two readers. To avoid interference from signals due to the shape of the reading area the reader was programmed to scan the centre part of the test strip.
Results Results for a panel of 20 samples are presented in Table 15. Discrepant results were obtained for several samples that gave a 1+ reading intensity for either visual or automatic reading.
Discussion Digital signal detection method allows objective signal quantification and has a strong correlation with the visual detection at a staining intensity ≥1+. Further software upgrading and standardization will be needed to allow fully automatic data processing and interpretation. Adjustment of the LFA cassette may be needed to flatten background levels currently affected by the presence of the notch dividing the reading window in a test and control zone and shadow arising at the left end of the test strip due to positioning of the light source. Removing of the notch may allow positioning of the test lines more centrally in the reading window away from the sample and conjugate pad in the current cassette and away from the shadow causing a back ground signal interfering with the HIV-1 test line.
HIV/Syphilis Combination LFA Assay using fluorescent labelled detection probes for automatic read out
Background The use of fluorescent labels instead of colloidal gold based nanoparticles could improve sensitivity and facilitate automatic reading. Colloidal gold nanoparticles are relatively unstable and the use of fluorescent labels could add to the robustness of the system. For fluorescent detection in LFA, Treponemal and HIV antigens were conjugated to a series of well-established commercial fluorescent labels including AlexaFluor-430, europium chelate reagents such as CM-EU and Dragon Green, and novel fluorescent polymers such as MEHPPV (Poly [2-methoxy-5-(2-ethylhexyloxy)-1, 4-phenyllenevinylene] nanoparticles developed by Dr. Mark Green at King’s College London) and BV605 (Brilliant Violet 605, Sirigen Inc. (Beckton, Dickinson & Company)). In initial investigations, fluorescent labels such as the MEHPPV, AlexaFluor-430 and Dragon Green gave either no or poor signal intensity, significant background activity and/or poor reproducibility when applied in LFA. More consistent results were obtained in LFA with CM-Europium (CM-EU) and Sirigen’s BV605 label conjugated to Treponemal specific antigen. Of these two labels, Sirigen’s BV605 showed relatively strong signals at the Treponemal test line when tested with syphilis positive serum samples (Figure 6, samples 1 to 10) and no reactivity with a set of control samples (Figure 6, samples 11 to 15). These results were obtained using the conditions summarized in Table 17a. Treponemal antigen labelled with CM-EU gave positive (1+ - 4+) results with all but one of the syphilis positive samples and weak positive (1+) results with the negative control samples (data not shown). Gp41 labelled with CM-EU for HIV-1 antibody detection showed inconsistent results with very strong signals for some HIV positive samples and no or very weak signals for others. Also a variety of staining artefacts including aspecific binding developed at the edges of the detection line for HIV negative samples was observed (data not shown) that was not solved by testing a range of conditions. Therefore the use of this label was not further investigated and further work concentrated on probes labelled with Sirigen’s BV605.
Method To circumvent some problems such as poor solubility of the detection probe when using the Sirigen’s BV605 in LFA, tests were performed in a “dipstick” format with the test strip placed in a reaction vessel containing a mixture of 2µl detection probe, 5µl serum and 200µl assay fluid. Assays were allowed to develop for 20min after which the results is documented using the CCM reader. Probes and assay conditions were initially developed for using Treponemal probes. To include HIV-1 and HIV-2 detection, the GP36/41 HIV-1/-2 fusion protein was included as second capture probe in a test line next to the Treponemal test line. This capture probes functions as capture probe for an HIV-1 antibodies and HIV-2 antibodies combined in a single test line. To minimize cross-reactivity at the Treponemal test line, the concentration and ratio of the TpN17/TpN47 capture proteins were adjusted and a different nitrocellulose detection strip was used (see Table 17b). The test was performed on 43 samples of which 22 were HIV1 and or HIV2 positive, 12 were positive for syphilis and 9 samples were negative to both syphilis and HIV1/2.
Results The sensitivity of the HIV/Syphilis test was 100% for HIV1/2 samples and 91% for syphilis positive samples. The specificity was 85% and 93%, respectively. Examples of HIV-1, HIV-2 and syphilis positive and negative test results are presented in Figure 7. Results of a direct comparison of LFA with fluorescent reading with visual detection in Prototype 2 & 3 LFAs using colloidal gold labelled detection probes is presented in Table 18.
Discussion Multiplexing for combined HIV-1/2 and syphilis antibody detection required major adjustments in the composition of the assay fluid compared with conditions established for each of the two systems separately which led to a decrease in signal intensity and specificity. Despite the relatively weak signal intensity the sensitivity was relatively high. The specificity in particular was low for HIV testing. The use of fluorescent detection seems to require more stringent conditions than the use of colloidal gold nanoparticles, limiting options for multiplexing, and optimized conditions negatively affected signal intensity and specificity. With the BV605 fluorescent-labelled probes some positive patient specimens with similar EIA titres and giving a similar staining intensity in LFA with gold-labelled probes gave strong differences in signal intensity. This suggests that whilst showing promise due to the brightness of the fluorophore, the biophysical characteristics of the BV605 fluorescent conjugates would require further optimisation to meet the specific requirements for this challenging HIV/syphilis LFA application.
Development of the HIV/Syphilis Combination Cyclotek Assay
Based on experience gained with the development of a multi-analyte LFA for HIV/Syphilis antigens and probes were selected for use in the development of the Cyclotek system. The Cyclotek systems differs from the immunochromatographic LFA in several aspects and one of these is the use of a solid support on the surface of which the capture probe is immobilized and the reaction takes place. In addition to different matrix types, gamma irradiation and pre-treatment of the matrix with corrosive agents and the type and composition of the binding fluid has been investigated to optimize binding of the capture probe. The use of a solid matrix also required the development and optimisation of a new assay buffer to dilute the sample, dissolve the detection probe and un the assay.. Antigens used as capture and detection probes were GP41 recombinant antigen, and TpN17 and TpN47 recombinant antigens at a ratio of 7:3. Both TpN17 and TpN47 as plain antigen and after conjugation to BSA. The BSA conjugates increased the appearance of the stained detection spots but also reduced the signal intensity and their use was not further investigated. Examples of buffers tested in the development of a HIV-1/Syphilis Cyclotek system are listed in Table 19. Buffer 1 (1xPBS pH 7.2 supplemented with 3% Tween-20, and 1.67% BSA) gave optimal functioning of the TpN-17 capture and detection probes and buffer 14 (1% Borax pH 7.0 supplemented with 1.5% Tween 20, 1% NaCl) gave promising results for the HIV-1 capture and detection probes. Buffers 1 and 14 differ significantly in composition and while buffer 1 gave poor results for HIV-1 detection, buffer 14 gave poor results for syphilis antibody detection. Based on the experience with lateral flow a combination of the two buffers (buffer system 4) was used to combine the two detection systems. The resulting buffer was tested at a wide pH range with a panel of serum samples used for test optimization. Combining the two buffers reduced signal intensity compared with buffers 1 and 14. The strongest signal intensity for the two analytes was obtained for buffer 4 adjusted to pH7.2 and buffer 4 with pH7.2 was therefore used for further test evaluation.
Test procedure The reaction in Prototype Cyclotek assay is started by the application of 175µl reaction mixture consisting of 150µl assay fluid (1xPBS, 0.5% borax, 2.25% Tween 20, 0.84% BSA, 0.5% NaCl pH7.2) 5µl detection reagent (GP160 4,5mg/l and TpN17 9mg/l conjugated to 40nm colloid gold, OD20) and 20µl sample onto the PE test surface of the Cyclotek. Tests were rotated for 15 min at a rotation speed of 5rpm. After 15 min the reaction mixture was removed and replaced with 150µl assay fluid and the rotation of the test disk was continued for 1 min. After removal of the test fluid the tests were examined by visual inspection for the presence of staining at the test dots (Figure 8). Tests were considered positive for HIV-1 when staining was observed at the HIV-1 test line and considered positive for syphilis when staining was observed at the Treponemal test line. Staining intensity at the test lines was rated in comparison with a scoring card with clearly visible staining scored 1+ to 4+ and considered positive, and very weak staining rated 0.5+ (doubtful) and considered negative. The rapid test was performed on 52 HIV positive and 49 syphilis positive samples obtained from BBI Solutions (Cardiff, UK). Information on the syphilis status was not available for the HIV positive serum samples and information on the HIV status was not available for the syphilis positive samples. After reading of the test result by visual inspection, disks were read digitally in the CCM reader.
Test results The main limitations of the optimized Cyclotek system are the low signal intensity obtained after 75 rotation rounds (15 min reaction time) and the presence of non-specific activity (Table 20). The signal intensity at the HIV detection spot was in general lower than the signal intensity at the Treponemal test spot and resulted in a relatively poor sensitivity for HIV (84.5%) compared with 100% for Treponemal antibody detection. Possibly the degree of false positive results (85.7% and 86.5% specificity, respectively) is overestimated as the HIV status of the syphilis positive control samples and the syphilis status of the HIV positive control samples was not available.
Conclusion Further options for optimization of the test conditions are limited due to the specific buffer requirements of two analyte systems but might be found in the use of another solid support and or the use of alternative detection reagents and blocking reagents.
HIV/Syphilis Combination Cyclotek Assay with automatic read out
A reader was constructed consisting of an AVT Mako G-125C color camera in combination with a 25mm Techspec Compact fixed focal lens. Dependent on the selected settings, the sample is illuminated with violet or white light (Figure 10). A simple automatic detection algorithm consisting of an image pre-processing and spot detection block is used to detect spots. For red spot detection created by colloidal gold labelled detection probes under white light illumination the difference between the red and green signal intensity is enhanced resulting in a better observable spot. Results of automatic reading was compared with visual reading of wet spots observed directly after the end of the incubation and washing procedure (1) and with visual reading after the remaining liquid left after the final washing step of the test procedure was dried (2). Automatic reading was performed on the dried disk with optimal reading settings for HIV and syphilis detection (3). However, even with these optimal settings, visual read out by method 1 and 2 result in a better accuracy than method 3 (Figure 11).
Discussion and Summary
A multi-analyte LFA was developed for the serodiagnosis of HIV and syphilis. The HIV1,2/syphilis LFA prototype 4 had a sensitivity of 98.8% for HIV-1, 100% for HIV-2 and 72.7% for syphilis at a specificity of 97.7%, 100% and 90.9%, respectively, and the sensitivity for VDRL test positive samples classified as lues latens/after treatment/long existing lues was 84.2% and 100% for active syphilis/recently treated. The diagnostic accuracy of the combination assay for HIV was similar to that of existing HIV/Syphilis POC tests.1,2 An asset of our HIV1,2/Syphilis LFA is that it allows discrimination between HIV-1 and HIV-2 infections. The diagnostic accuracy of the HIV1,2/Syphilis LFA for the serodiagnosis of syphilis may need further fine-tuning as in different prototypes the sensitivity and specificity varied and were somewhat lower than reported for existing LFAs. In published studies the Treponemal Pallidum Particle Agglutination test was used as comparator and whereas we used a highly sensitive EIA3 as reference, this could explain the observed lower sensitivity for our LFA. The majority of the LFA negative test results obtained for EIA positive samples had relatively low EIA values <7.0 that tested lues blot positive but VDRL test negative and that were classified as earlier infection after treatment/lues latens if never treated. A further independent laboratory based evaluation study is ongoing to confirm the diagnostic performance of the HIV1,2/Syphilis LFA prototype 4. The standard operation procedures for the HIV1,2/Syphilis LFA Prototype 4 have been documented and a manufacturing file is available.
Fluorescent labels were tested to determine whether the use of fluorescent detection probes could improve test performance. Of the different fluorescent probes tested Sirigen’s BV605 gave the most promising results in LFA and a HIV/Syphilis dipstick was developed with a sensitivity of 100% for HIV1/2 samples and of 91% for syphilis positive samples. The specificity was 85% and 93%, respectively. In this dipstick test liquid detection probe was used, and for conversion to LFA it would be necessary to establish conditions for drying and conservation of the fluorescent labelled probes onto the conjugate pad of the LFA device and that allow rapid release of functional detection reagent by the application of the sample and assay fluid mixture. Whereas Sirigen’s BV605 provided a strong and specific signal the inability to identify a common assay fluid for the optimal serodiagnosis of HIV and syphilis in one assay limited its use in the development of a HIV/syphilis combination assay.
Options for optimization of the HIV1,2/Syphilis LFA test performance were restricted by the differences in requirements in buffer and buffer additives for the optimal functioning of the two detection systems. Requirements for the optimal performance of the two detection systems in the Cyclotek system appeared to be even more sensitive to the choice of assay components and reagents and the composition of the assay fluid thus further restricted options for optimization. A major factor that hindered test development was the poor binding of the capture probes to the solid support. Whereas relatively strong signals were obtained for syphilis positive control samples, weaker signals were obtained for HIV positive control samples and the signal strength at the HIV capture probe would need to be improved to improve specificity. Automatic read-out of Cyclotek test signals did contribute to standardized test interpretation but did not improve sensitivity.
The use of existing readers developed for LFA such as the Fio Deki reader may facilitate standardized reading of the HIV/Syphilis LFA developed here.5,6 The use of a reader may also improve sensitivity as is observed for the reading of malaria POC testing.
Literature references
1. Bristow CC, Adu-Sarkodie Y, Ondondo RO et al. Multisite Laboratory Evaluation of a Dual Human Immunodeficiency Virus (HIV)/Syphilis Point-of-Care Rapid Test for Simultaneous Detection of HIV and Syphilis Infection. Open Forum Infect Dis. 2014;1:ofu015.
2. Yin YP, Ngige E, Anyaike C et al. Laboratory evaluation of three dual rapid diagnostic tests for HIV and syphilis in China and Nigeria. Int J Gynaecol Obstet. 2015;Suppl 1:S22-6.
3. Knight CS1, Crum MA, Hardy RW. Evaluation of the LIAISON chemiluminescence immunoassay for diagnosis of syphilis. Clin Vaccine Immunol. 2007;14:710-3.
4. Causer LM, Kaldor JM, Fairley CK et al. A laboratory-based evaluation of four rapid point-of-care tests for syphilis. PLoS One. 2014 Mar 11;9(3):e91504.
5. Shekalaghe S, Cancino M, Mavere C, Juma O, Mohammed A, Abdulla S, Ferro S. Clinical performance of an automated reader in interpreting malaria rapid diagnostic tests in Tanzania. Malar J. 2013;12:141.
6. Herrera S, Vallejo AF, Quintero JP et al. Field evaluation of an automated RDT reader and data management device for Plasmodium falciparum/Plasmodium vivax malaria in endemic areas of Colombia. Malar J. 2014;13:87.
Potential Impact:
Due to limited awareness among pregnant women and health care providers, lack of availability of simple and rapid diagnostics tests that combine HIV and syphilis detection and lack of access to diagnostic facilities where diagnostic testing for HIV and syphilis may be performed many pregnant women at risk of infection with either or both of the pathogens do not receive an appropriate diagnosis and are not treated. As a result, the risk of congenital transmission with severe damage to the child remains high. The wider availability of affordable high quality point-of-care diagnostics will contribute to improved awareness and a better diagnosis and treatment. The development of the HIV1,2/Syphilis LFA developed by the PreventIt team is the first step towards a better access of pregnant women to diagnosis and treatment for HIV and syphilis. Unfortunately, due to a management decision to stop all laboratory activities of partner 2 the diagnostic performance of the HIV1,2/Syphilis LFA could not be confirmed in a field study by testing pregnant women at risk of HIV and Syphilis infection attending a mother and child clinic in sub-saharan Africa. As a result, essential data is lacking to bring the test to the market. Some of the laboratory activities of partner 1 on point of care test development will be taken over by the newly established Mondial Diagnostics Foundation. This foundation was established to develop test and to bring test more efficiently to the market and to serve patients more effectively.