| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
 |
ABSTRACT |
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The most studied malaria RDTs offer simple identification of two parasite antigens: histidine-rich protein 2 (HRP2) and plasmodium lactate dehydrogenase (pLDH). HRP2 was the first antigen targeted by an RDT,9 has been available in various commercial formats for several years, has shown good sensitivity in a variety of field settings, and is increasingly advocated as a diagnostic test where reliable microscopy is not available. A potential problem for HRP2-based assays is persistence of detectable circulating antigen for up to several weeks after parasites have been eradicated.10–12 Persistent HRP2 antigenemia has not correlated with treatment failure, suggesting that this finding is not representative of persistent infection.10,12 Persistent antigenemia thus may limit the usefulness of HRP2-based assays in areas of intense malaria transmission, where positive tests may commonly be due to prior infections that are no longer clinically relevant. pLDH-based RDTs appear to be slightly less sensitive than those detecting HRP2, but the antigen is rapidly cleared from the bloodstream, becoming undetectable at about the same time blood smears become negative after antimalarial therapy.13–15 Thus, if sensitivity is adequate, the increased specificity of pLDH-based assays for acute malaria suggests that they may be better-suited for high-transmission areas, such as much of sub-Saharan Africa. With increasing advocacy for the implementation of RDTs, it is critical that optimal diagnostic strategies are identified. The true impact of the varied sensitivity and specificity of different tests is best compared with long-term follow-up to consider the impacts of prior infections and persistent antigenemia on test results. For this reason, we compared the diagnostic accuracy of HRP2- and pLDH-based RDTs, using expert microscopy as the gold standard, in a longitudinal cohort of children in Kampala, Uganda.
|
METHODS |
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
38°C or history of fever within the previous 24 hours), a fingerprick blood sample is obtained for thick and thin smears and storage on filter paper. If the blood smear is positive, the child is treated with antimalarials and followed for 28 days; if the smear is negative, the child does not receive antimalarials and is treated according to standard clinical algorithms and the study physician’s judgment. Parents/guardians gave informed consent for all study procedures, and the study was approved by the Uganda National Council of Science & Technology and by the institutional review boards of Makerere University and the University of California, San Francisco.
RDT study methods.
At the time of the RDT evaluation, children in the cohort ranged in age from 1.5 to 11.5 years. From October 2005 to May 2006, each time a blood smear was done to evaluate fever in a study participant, except when the fever occurred within 3 days of a confirmed episode of malaria, a fingerprick blood sample was obtained for the two RDTs in addition to thick and thin smears and storage on filter paper (Figure 1
). If a participant presented with repeated episodes of fever after diagnosis of a non-malarial illness, the RDT was repeated at the study physician’s discretion. Clinical management was guided by microscopy results; RDT results did not influence patient care.
|
2000/µL, the final density recorded was that of the third reader. RDTs were selected for evaluation on the basis of ease of use (relatively few preparation steps and clear distinction between positive and negative results), safety (minimal exposure to blood during test preparation), completeness of packaging and labeling, appropriate packaging for transport and storage in tropical environments (each test individually wrapped in foil with plastic liner and desiccant), low market price, and reliability of supply. The RDTs studied were Paracheck (detection of HRP2, Orchid Biomedical Systems, Goa, India) and Parabank (detection of pLDH, Zephyr Biomedicals, Goa, India). RDTs were obtained directly from the manufacturers and stored in their original packaging at room temperature in the study clinic. Temperature and humidity of the storage area were monitored, but not controlled. Over the course of the study period, the temperature in the storage area ranged from 19 to 29°C, with a mean low of 24°C and a mean high of 27°C. The relative humidity ranged from 31% to 82%, with a mean low of 53% and a mean high of 69%. Prior to the beginning of the study, positive and negative control blood samples were obtained, and stored at –80°C for quality-control testing of RDTs throughout the study. Each batch of RDTs underwent quality-control testing when opened and at 8- to 12-week intervals thereafter. The two positive control samples had densities 84/µL and 5000/µL, respectively. All HRP2 RDTs tested with quality-control samples were accurate; all pLDH RDTs tested were accurate for the negative and 5000/µL samples, but only 2 of 8 were accurate for the 84/µL sample.
RDTs were prepared and read by study physicians and then read by laboratory technicians. All readers were trained to perform the tests according to manufacturers’ instructions. Study physicians interpreted and recorded RDT results after 15 minutes, at which time they were unaware of blood smear results. They were advised that if the background of the RDT test window remained pink (bloody) at the end of 15 minutes, they should allow the background to clear before reading the RDT. RDTs were then carried to the adjacent study laboratory, where they were re-read by laboratory technicians who were unaware of both the physician’s interpretation and the patient’s microscopy result. Readers recorded RDT results as either positive or negative; they were trained to consider faint test lines as positive.
Molecular methods. PCR was performed to identify parasite species in samples positive by microscopy but negative by RDT, as well as to detect subpatent infections in samples negative by microscopy but positive by RDT, and in a random sample of microscopy-negative and RDT-negative samples. DNA was extracted from filter paper samples using Chelex resin17 and stored at –20°C until use. To detect Plasmodium falciparum, the block-3 region of merozoite surface protein-2 (msp-2) was amplified by nested PCR with primers corresponding to conserved sequences flanking this region18 followed by primers to amplify the IC3D7 and FC27 allelic families, using conditions described previously.19 In addition, to detect P. falciparum, P. vivax, P. malariae, and P. ovale, genus-specific followed by nested species-specific PCR of 18S small subunit ribosomal RNA20 (ssu rRNA) for the four species (Malaria Research and Reference Reagent Resource Center, Manassas, VA) was performed, using oligonucleotide primers and conditions as described previously.21 PCR products were analyzed by electrophoresis using 2% agarose gels.
Statistical methods.
Data were entered using Epi-Info version 6.04 (Centers for Disease Control and Prevention, Atlanta, GA) and analyzed with Stata version 8.0 (Stata Corporation, College Station, TX). Sensitivity, specificity, and positive and negative predictive values were calculated by comparing the proportion of positive and negative results for each RDT with expert microscopy. Categorical variables were compared using 
2 or Fisher’s exact test. A P value of < 0.05 was considered statistically significant.
|
RESULTS |
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Over the 8-month period, 868 fevers were new fevers in a previously well child, 21 occurred 4–14 days after a diagnosis of malaria, and 44 occurred during follow-up of a non-malarial illness. RDTs were not performed in 15 episodes, in 9 at the discretion of the physician during follow-up of a non-malarial febrile illness, and in 6 because of protocol errors. Light microscopy identified positive malaria smears in 289 episodes (31%). Blood smear results served as the gold standard for comparison with RDT results. As RDT results are dependent on reader accuracy, we compared readings by two groups of clinic personnel: study physicians and laboratory technicians. In both cases, the sensitivity (> 92%) and negative predictive value (> 96%) were higher for the HRP2 assay, and specificity (> 99%) and positive predictive value (> 99%) were higher for the pLDH assay (Figure 2). First readers interpreted RDT results an average of 15 minutes after preparation, and second readers interpreted results an average of 7 minutes later. First and second test readings agreed in 98% of readings; they disagreed for 16 HRP2 tests and 13 pLDH tests. For 14/16 (88%) discordant HRP2 readings and 10/13 (77%) discordant pLDH readings, only second readings were positive.
|
Of the 22 false-negative HRP2 results (based on first reading), 15 (68%) occurred in non-falciparum infections (Figure 3). Of the remaining 7 false negatives, 5 were interpreted as positive by the second reader. The 2 remaining false negatives occurred in a P. falciparum mono-infection with parasite density 48/µL, and a P. falciparum and P. vivax mixed infection with density 680/µL. Considering only P. falciparum infections, the sensitivity of the HRP2 assay at the second reading was 99% (272/274).
|
1000/µL, respectively (P < 0.0001). Evaluation of false-positive results. Possible reasons for false-positive RDT results include persistent antigenemia after antimalarial treatment, detection of gametocytes when asexual forms are not present, RDT detection of low-density microscopy-negative infections, or presence of antigenemia early in infection before parasites are detected by microscopy.
Of the 42 false-positive HRP2 results, 12 (29%) occurred within 14 days of a prior diagnosis of malaria, 26 (62%) within 28 days, and 32 (76%) within 42 days. In contrast, negative HRP2 results occurred as early as 7 days after initial diagnosis of a previous episode of malaria.
Gametocytes were detected by microscopy in only 12 of the 918 cases (1.3%). No HRP2 result was positive in a case where the smear showed gametocytes but not asexual parasites.
PCR was conducted to assess whether false-positive RDT results may have been associated with subpatent parasitemia. Of 40 evaluable false-positive HRP2 results, PCR was positive for P. falciparum in 8 (20%), compared with PCR positivity in 5/66 (8%) of control HRP2- and microscopy-negative samples (P = 0.07). Four of the 8 smear-negative, RDT- and PCR-positive samples were obtained within 28 days of a prior episode of malaria.
Negative HRP2 results were recorded up to 3 days prior to an episode of malaria. Only one patient developed malaria within a week after a false-positive HRP2 result. The sample from the initial evaluation showed no asexual parasites or gametocytes but was positive for P. falciparum by PCR. The patient returned with persistent fever 5 days after initial evaluation, at which time the blood smear was positive, with parasite density 52,840/µL.
Only one pLDH test result was false-positive, in a patient who had no documented previous episode of malaria over 469 days of follow-up, and no malaria during the subsequent 2 months of study follow-up. No gametocytes were seen in the smear, the sample was negative by PCR for all four malaria species, and the second reading of the RDT was negative, strongly suggesting that this false positive was due to an error during the first test reading.
|
DISCUSSION |
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
In Uganda, RDTs are increasingly available in the private health care sector and are widely advocated for use in the public sector, though clear guidelines or algorithms for their use are lacking. In Kampala, both the HRP2 and pLDH tests showed a high negative predictive value and appeared to offer good reliability in ruling out malaria as the cause of a fever. Considering the potential values of RDTs, some limitations in both sensitivity and specificity may be acceptable. The lower specificity of the HRP2-based test, due to persistent antigen-emia after recent infections, may lead to some inappropriate treatments, but many fewer than if all episodes of fever were treated as malaria. However, the lower specificity of HRP2 assays may be more problematic, with many more unnecessary malaria treatments, in regions with higher transmission intensity than Kampala. The lower sensitivity of the pLDH-based assay might also be a concern, but in Kampala, missed episodes were primarily of relatively low parasitemia, suggesting that, in immune populations, mostly mild or asymptomatic infections will be missed. Indeed, especially if technological innovations can improve the sensitivity of pLDH-based tests, they may well offer the optimal balance of sensitivity and specificity for the diagnosis of malaria in Africa.
To our knowledge, this study offers the first comparison of RDTs in a longitudinal format, allowing assessment of the importance of previous and future malaria infections in RDT accuracy. A number of other RDT evaluations have been conducted, though results have varied widely, likely due at least in part to different methodologies and locations. Two previous RDT studies in western Uganda compared HRP2-based tests with expert microscopy. One evaluation, using an older HRP2 assay, found a sensitivity of 99.6% for parasitemia > 500/µL and specificity of 92.7% in patients with fever.24 The other study, using the same HRP2 test as in our evaluation, found a sensitivity of 97% and specificity of 88% for P. falciparum infections.25 These estimates are similar to those for the HRP2-based test in our current evaluation. Our results also confirm the superior specificity of pLDH seen in a study in Tanzania,26 although sensitivity of both tests was somewhat lower in our study.
Our results are not immediately applicable to fever case management across Africa. We obtained RDTs directly from manufacturers, and we used and stored kits as recommended by manufacturers; adherence to these guidelines may be challenging in rural settings, and test quality is likely to deteriorate if kits are less well maintained.27 Our evaluation was performed in an area with relatively low malaria transmission. Because of the location and design of our study, our patients likely presented to the clinic earlier in the course of malaria than in non-research settings. Our staff was carefully trained in use of the two RDTs before initiation of our study; test accuracy may be lower in field settings, although a number of reports indicate that village health workers with minimal training are able to satisfactorily prepare and interpret RDTs.28,29 Considering these limitations, how should the results of this evaluation influence malaria treatment policy? For Kampala, our results suggest that, in settings without access to microscopy, use of either HRP2- or pLDH-based RDTs could dramatically lower the use of inappropriate antimalarial therapy without missing many episodes of clinical malaria. However, it will be necessary to perform similar analyses in areas with different epidemiology to determine the predictive values of different RDTs in various settings. In addition, the issue of cost and cost-effectiveness of RDTs, compared with presumptive treatment and with diagnosis with microscopy, must be considered. In the era of artemisinin combination therapies, using RDTs to target treatment to confirmed cases of malaria may help to maximize the impact of these valuable resources.
Received January 12, 2007. Accepted for publication February 21, 2007.
Acknowledgments: The authors thank the study participants and their families, and Makerere University–University of California, San Francisco (MU-UCSF) Malaria Research Collaboration clinic and laboratory staff, including Regina Nakafeero, Maxwell Kilama, Christopher Bongole, Felix Jurua, Irene Namukwaya, Denise Njama-Meya, Bridget Nzarubara, Catherine Maiteki, Joy Bossa, Ruth Namuyinga, Arthur Mpimbaza, Joanita Nankabirwa, Florence Nankya, and Moses Musinguzi. We are grateful to Sam Nsobya, Moses Kiggundu, and Chris Dokomajilar for assistance with molecular studies, to John Patrick Mpindi and Geoff Lavoy for assistance with computer maintenance and data management, and to Catherine Tugaineyo, Richard Oluga, Kenneth Mwebaze, and Peter Padilla for administrative support. The authors thank Carole Fogg and Patrice Piola of MSF/Epicentre for providing preliminary data from an RDT study in Mbarara, western Uganda. We also thank the U.S. National Institutes of Health and the Doris Duke Charitable Foundation for their support. P.J.R. is a Doris Duke Charitable Foundation Distinguished Clinical Scientist.
Financial support: This study was funded by the U.S. National Institutes of Health (grant K23 AI065457-01), with additional support received from the Doris Duke Charitable Foundation.
* Address correspondence to H. Hopkins, c/o Makerere University–University of California, San Francisco, Malaria Research Collaboration, P.O. Box 7475, Kampala, Uganda. E-mail: hhopkins{at}medsfgh.ucsf.edu 
Authors’ addresses: Heidi Hopkins, University of California, San Francisco, c/o MU-UCSF Malaria Research Collaboration, P.O. Box 7475, Kampala, Uganda, Telephone/Fax: +256-414-540524, E-mail: hhopkins{at}medsfgh.ucsf.edu. Kambale Wilson, MU-UCSF Malaria Research Collaboration, P.O. Box 7475, Kampala, Uganda, Telephone: +256-414-530692, Fax: +256-414-540524. Moses R. Kamya, Makerere University Medical School, P.O. Box 7072, Kampala, Uganda, Telephone: +256-414-541188, Fax: +256-414-540524. Sarah Staedke, London School of Hygiene & Tropical Medicine, c/o MU-UCSF Malaria Research Collaboration, P.O. Box 7475, Kampala, Uganda, Telephone: +256-414-530692, Fax: +256-414-540524. Grant Dorsey and Philip J. Rosenthal, University of California, San Francisco, Parnassus Avenue, Box 0811, San Francisco, CA 94143, Telephone: +1 (415) 648-4680, Fax: +1 (415) 648-8425.
Reprint requests: H. Hopkins, c/o Makerere University–University of California, San Francisco Malaria Research Collaboration, P.O. Box 7475, Kampala, Uganda, Telephone/Fax: +256-414-540524, E-mail: hhopkins{at}medsfgh.ucsf.edu.
|
REFERENCES |
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. Skarbinski, P. O. Ouma, L. M. Causer, S. K. Kariuki, J. W. Barnwell, J. A. Alaii, A. M. de Oliveira, D. Zurovac, B. A. Larson, R. W. Snow, et al. Effect of Malaria Rapid Diagnostic Tests on the Management of Uncomplicated Malaria with Artemether-Lumefantrine in Kenya: A Cluster Randomized Trial Am J Trop Med Hyg, June 1, 2009; 80(6): 919 - 926. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. de Oliveira, J. Skarbinski, P. O. Ouma, S. Kariuki, J. W. Barnwell, K. Otieno, P. Onyona, L. M. Causer, K. F. Laserson, W. S. Akhwale, et al. Performance of Malaria Rapid Diagnostic Tests as Part of Routine Malaria Case Management in Kenya Am J Trop Med Hyg, March 1, 2009; 80(3): 470 - 474. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Ratsimbasoa, L. Fanazava, R. Radrianjafy, J. Ramilijaona, H. Rafanomezantsoa, and D. Menard Evaluation of Two New Immunochromatographic Assays for Diagnosis of Malaria Am J Trop Med Hyg, November 1, 2008; 79(5): 670 - 672. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. M. Kifude, H. G. Rajasekariah, D. J. Sullivan Jr., V. A. Stewart, E. Angov, S. K. Martin, C. L. Diggs, and J. N. Waitumbi Enzyme-Linked Immunosorbent Assay for Detection of Plasmodium falciparum Histidine-Rich Protein 2 in Blood, Plasma, and Serum Clin. Vaccine Immunol., June 1, 2008; 15(6): 1012 - 1018. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. O. Wilson, A. A. Adjei, W. Anderson, S. Baidoo, and J. K. Stiles Detection of Plasmodium falciparum Histidine-rich Protein II in Saliva of Malaria Patients Am J Trop Med Hyg, May 1, 2008; 78(5): 733 - 735. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
