The study was performed in twelve, not adjacent, commune health posts and the out patient clinic of the provincial malaria station of Binh Thuan, a province in southern Vietnam. Patients presenting with acute undifferentiated fever (AUF) at these primary health care facilities were included in this study. AUF was defined as any febrile illness of duration less than 14 days, confirmed by an axillary temperature ≥ 38.0°C, without any indication for either severe systemic or organ specific disease. Malaria was excluded by microscopic examination of a thick blood smear.
Record forms were filled out for all AUF patients, recording patient identifiers, duration of disease, signs and symptoms. Data and blood were collected on presentation (t0) and all included subjects were asked to come back after 3 weeks (t3) for re-assessment and collection of a second blood sample.
Collection and storage of blood samples
Blood was collected by venapuncture. A few circles of 15 mm, printed on specimen collection paper (cotton linters paper, S&S 903, Schleicher & Shuell, Dassel, Germany), was filled with full blood and left to dry at the air, avoiding exposure to sunlight. They were packed separately in zip-closure nylon bags, with a few silica grains to prevent moisturizing. The remaining blood in the collection tube was left to clot and thereafter centrifuged at 1000 rpm for 15 minutes at the spot. The serum was decanted into two polypropylene vials and stored in a freezer at minus 20°C at the health facilities. The samples were collected once a month by the research team of Cho Ray Hospital and transported to Ho Chi Minh City in a cooling box, filled with ice packs (duration of this travel is approximately 4 hours). In Ho Chi Minh City the serum samples were stored at minus 70°C until analysis. The filter papers were kept in their nylon bags in an air-conditioned room.
Reconstitution of sample from filter papers
Samples from dried blood filter paper were reconstituted as described previously In brief, two 6 mm disks were punched from the filter paper and soaked in a vial over night at 4°C on an automatic shaker with 0.2 ml of phosphate-buffered saline supplemented with 2% fetal bovine serum. The following day the supernatant was stirred, collected and analyzed, together with the corresponding serum pairs. The resulting sample was treated as a 1:20 dilution of human serum, assuming that the two circles contain approximately 25 μL of blood, corresponding to 10 μl of serum.
IgM and IgG of both t0 and t3 specimens were always tested in the same batch, on the same day. All samples were tested in the laboratory of microbiology of Cho Ray Hospital, Ho Chi Minh City (CRH), Vietnam. A subset of sera was also tested in the Laboratory of Virology, Erasmus Medical Center, Rotterdam, the Netherlands (EMC), by the same persons, following the same protocol and using similar test kits with the same batch numbers. The concentrations of dengue-virus specific IgM antibodies were measured using a commercially available IgM capture ELISA (Dengue Fever Virus IgM Capture ELISA, Focus diagnostics, Cypress, CA, USA), as described previously[3, 4] For IgG antibodies an indirect ELISA was used (Dengue Fever Virus IgG Focus diagnostics, Cypress, CA, USA). Both the IgM and IgG ELISAs detect antibodies against an antigen mixture of equal portions of DEN 1–4. The ELISA plates were read by optical density (OD) ELISA plate readers. OD values were measured at 450 nm with 620 nm as a reference. In CRH a Benchmark microplate reader was used (Bio-Rad Laboratories, Inc., Hercules, CA, USA), in the EMC an Anthos Reader 2001 Type 10 500 (Anthos Labtec Instruments, Salzburg, Austria). Serum and the reconstituted filter paper samples, both IgG and IgM, were tested at a final dilution of 1:100 (implying an extra 1: 5 dilution of the filter paper supernatant). Following the instructions of the manufacturer, the OD results were corrected by subtracting the OD values of blank samples which are included in every test kit. The ratio with a standard positive sample, the cut-off calibrator, is expressed as an index value (IV). All samples were tested in duplicate and the mean IVs were used for further analysis. Quality control is provided by testing three control sera, enclosed in every test kit. IV < 1 is considered negative, IV > 1 positive. The value 1 is not specified. In this study we chose to interpret an IV of exactly 1 as positive.
In order to analyze misclassification and agreement, the results were categorized into four diagnoses. A molar increase of dengue antibody concentrations, measured by ELISA, is generally regarded significant In this study a four fold increase of IV was taken as a significant, based on a previous study by Cobelens The IgM concentration on t3, relative to the IgG concentration on t3 was also used as a criterion, recognizing that this ratio between the results of a capture IgM ELISA and direct IgG ELISA is slightly different from the method published by the WHO. Acute primary dengue virus infection was defined as a positive IgM on t3 with an IgM/IgG ratio on t3 greater than one. A positive IgM on t3 with an IgM/IgG ratio on t3 less then one, or a negative IgM reaction on t3 but with a positive IgGt3 and a fourfold molar increase of IgG between t0 to t3 were classified as acute secondary dengue. A negative IgM reaction on t3, a positive IgG on t3 but without a fourfold increase between t0 and t3 was classified as "not acute dengue but past infection", and a case of both negative IgM and IgG on t3 was classified as "no dengue".
Intra-individual variation of IgG IVs was investigated in the t0 and t3 sera of subjects classified as cases of past dengue or no dengue. ELISA results of serum and filter paper samples were compared by linear regression and calculation of the coefficient correlation. Agreement between the tests done in the two different laboratories and between serum and filter paper results were analyzed by comparing the positive and negative results and the diagnostic classification based on paired serum samples. The decay during storage of serum was analyzed with a linear mixed effects model of IV as dependent variable and the time between sampling and testing blood, t-store, as independent variable, applied to the sera which were tested twice. The decay of filter papers was estimated by ordinary least squares linear regression of the difference between the serum and corresponding filter paper IVs, ΔIV, against t-store. Statistical analysis was done using SPSS (version 11.5, SPSS Inc. Ill.). We choose to present scatter plots, instead of Bland Altman plots, because that also illustrates the distinction between positive and negative results.
The study was approved by the Review Board of the Cho Ray Hospital, Ho Chi Minh City, Vietnam. The study was explained and discussed in meetings with provincial authorities and health post staffs. All patients, (or their parents or guardian) gave their written informed consent (for children from parent or guardian).