Journal Article > ResearchFull Text
J Acquir Immune Defic Syndr. 25 July 2013; Volume 64 (Issue 2); DOI:10.1097/QAI.0b013e3182a61e63
Pannus P, Fajardo E, Metcalf CJ, Coulborn R M, Duran LT, et al.
J Acquir Immune Defic Syndr. 25 July 2013; Volume 64 (Issue 2); DOI:10.1097/QAI.0b013e3182a61e63
: Rollout of routine HIV-1 viral load monitoring is hampered by high costs and logistical difficulties associated with sample collection and transport. New strategies are needed to overcome these constraints. Dried blood spots from finger pricks have been shown to be more practical than the use of plasma specimens, and pooling strategies using plasma specimens have been demonstrated to be an efficient method to reduce costs. This study found that combination of finger-prick dried blood spots and a pooling strategy is a feasible and efficient option to reduce costs, while maintaining accuracy in the context of a district hospital in Malawi.
Journal Article > ResearchFull Text
Trans R Soc Trop Med Hyg. 31 January 2008
Zachariah R, Harries AD, Spielmann M P, Arendt V, Nchingula D, et al.
Trans R Soc Trop Med Hyg. 31 January 2008
In Thyolo district, Malawi, an operational research study is being conducted on the efficacy and feasibility of co-trimoxazole prophylaxis in preventing deaths in HIV-positive patients with tuberculosis (TB). A series of cross-sectional studies were carried out in 1999 and 2001 to determine (i) whether faecal Escherichia coli resistance to co-trimoxazole in TB patients changed with time, and (ii) whether the resistance pattern was different in HIV-positive TB patients who were taking co-trimoxazole prophylaxis. Co-trimoxazole resistance among E. coli isolates in TB patients at the time of registration was 60% in 1999 and 77% in 2001 (P < 0.01). Resistance was 89% among HIV-infected TB patients (receiving cotrimoxazole), while in HIV-negative patients (receiving anti-TB therapy alone) it was 62% (P < 0.001). The study shows a significant increase of E. coli resistance to co-trimoxazole in TB patients which is particularly prominent in HIV-infected patients on co-trimoxazole prophylaxis. Since a high degree of plasmid-mediated transfer of resistance exists between E. coli and the Salmonella species, these findings could herald limitations on the short- and long-term benefits to be expected from the use of co-trimoxazole prophylaxis in preventing non-typhoid Salmonella bacteraemia and enteritis in HIV-infected TB patients in Malawi.
Journal Article > ResearchAbstract
Trop Med Int Health. 7 November 2020; Volume 26; DOI:10.1111/tmi.13519
Gueguen M, Nicholas S, Poulet E, Schramm B, Szumilin E, et al.
Trop Med Int Health. 7 November 2020; Volume 26; DOI:10.1111/tmi.13519
Objective
We monitored a large‐scale implementation of the Simple Amplification‐Based Assay semi‐quantitative viral load test for HIV‐1 version I (SAMBA I Viral Load = SAMBA I VL) within Médecins Sans Frontières’ HIV programmes in Malawi and Uganda, to assess its performance and operational feasibility.
Methods
Descriptive analysis of routine programme data between August 2013 and December 2016. The dataset included samples collected for VL monitoring and tested using SAMBA I VL in five HIV clinics in Malawi (four peripheral health centres and one district hospital), and one HIV clinic in a regional referral hospital in Uganda. SAMBA I VL was used for VL testing in patients who had been receiving ART for between 6 months and ten years, to determine whether plasma VL was above or below 1000 copies/mL of HIV‐1, reflecting ART failure or efficacy. Randomly selected samples were quantified with commercial VL assays. SAMBA I instruments and test performance, site throughput, and delays in communicating results to clinicians and patients were monitored.
Results
Between August 2013 and December 2016 a total of 60 889 patient samples were analysed with SAMBA I VL. Overall, 0.23% of initial SAMBA I VL results were invalid; this was reduced to 0.04% after repeating the test once. Global test failure, including instrument failure, was 1.34%. Concordance with reference quantitative testing of VL was 2620/2727, 96.0% (1338/1382, 96.8% in Malawi; 1282/1345, 95.3% in Uganda). For Chiradzulu peripheral health centres and Arua Hospital HIV clinic, where testing was performed on‐site, same‐day results were communicated to clinicians for between 91% and 97% of samples. Same‐day clinical review was obtained for 84.7% across the whole set of samples tested.
Conclusions
SAMBA I VL testing is feasible for monitoring cohorts of 1000 to 5000 ART‐experienced patients. Same‐day results can be used to inform rapid clinical decision‐making at rural and remote health facilities, potentially reducing time available for development of resistance and conceivably helping to reduce morbidity and mortality.
We monitored a large‐scale implementation of the Simple Amplification‐Based Assay semi‐quantitative viral load test for HIV‐1 version I (SAMBA I Viral Load = SAMBA I VL) within Médecins Sans Frontières’ HIV programmes in Malawi and Uganda, to assess its performance and operational feasibility.
Methods
Descriptive analysis of routine programme data between August 2013 and December 2016. The dataset included samples collected for VL monitoring and tested using SAMBA I VL in five HIV clinics in Malawi (four peripheral health centres and one district hospital), and one HIV clinic in a regional referral hospital in Uganda. SAMBA I VL was used for VL testing in patients who had been receiving ART for between 6 months and ten years, to determine whether plasma VL was above or below 1000 copies/mL of HIV‐1, reflecting ART failure or efficacy. Randomly selected samples were quantified with commercial VL assays. SAMBA I instruments and test performance, site throughput, and delays in communicating results to clinicians and patients were monitored.
Results
Between August 2013 and December 2016 a total of 60 889 patient samples were analysed with SAMBA I VL. Overall, 0.23% of initial SAMBA I VL results were invalid; this was reduced to 0.04% after repeating the test once. Global test failure, including instrument failure, was 1.34%. Concordance with reference quantitative testing of VL was 2620/2727, 96.0% (1338/1382, 96.8% in Malawi; 1282/1345, 95.3% in Uganda). For Chiradzulu peripheral health centres and Arua Hospital HIV clinic, where testing was performed on‐site, same‐day results were communicated to clinicians for between 91% and 97% of samples. Same‐day clinical review was obtained for 84.7% across the whole set of samples tested.
Conclusions
SAMBA I VL testing is feasible for monitoring cohorts of 1000 to 5000 ART‐experienced patients. Same‐day results can be used to inform rapid clinical decision‐making at rural and remote health facilities, potentially reducing time available for development of resistance and conceivably helping to reduce morbidity and mortality.
Journal Article > ResearchFull Text
J Clin Microbiol. 1 May 2014; Volume 52 (Issue 5); 1343-1351.; DOI:10.1128/JCM.03519-13
Farjardo E, Metcalf CJ, Chaillet P, Aleixo L, Pannus P, et al.
J Clin Microbiol. 1 May 2014; Volume 52 (Issue 5); 1343-1351.; DOI:10.1128/JCM.03519-13
HIV-1 viral load (VL) testing is not widely available in resource-limited settings. Use of finger-prick dried blood spot (FP-DBS) samples could remove barriers related to sample collection and transport. Measurement of VL using DBS from EDTA venous blood (VB-DBS) in place of plasma has previously been validated using the NucliSENS EasyQ HIV-1 v2.0 assay, but information on the accuracy of FP-DBS samples for measuring VL is limited. This prospective study, conducted at Thyolo District Hospital in Southern Malawi, compared VL levels measured on FP-DBS samples and plasma, using the NucliSENS EasyQ HIV-1 v2.0 assay. Comparability was assessed by means of agreement and correlation (131 patients with VLs ≥100 copies/ml), and sensitivity and specificity (612 patients on ART). Samples of EDTA venous blood and FP-DBS from 1,009 HIV-infected individuals were collected and prepared in the laboratory. Bland-Altman analysis found good agreement between plasma and FP-DBS VL levels, with a mean difference of -0.35 log10, and 95% limits of agreement from -1.26 to 0.55 log10. FP-DBS had a sensitivity of 88.7% (95% confidence interval [CI]: 81.1 - 94.4%) and specificity of 97.8% (95% CI: 96.1 - 98.9%) using a 1,000 copies/ml cut-point; and a sensitivity of 83.0% (95% CI: 73.4 - 90.1%) and specificity of 100% (95% CI: 99.3-100%) using a 5,000 copies/ml cut-point. This study shows that FP-DBS is an acceptable alternative to plasma for measuring VL using the NucliSENS EasyQ HIV-1 v2.0. We are conducting a second study to assess the proficiency of health workers at preparing FP-DBS in primary healthcare clinics.