The use of the single radial haemolysis technique in the serological diagnosis of dengue and Japanese encephalitis virus infections*

Bulletin of the World Health Organization, 63 (6): 1043-1053 (1985) C World Health Organization 1985 The use of the single radial haemolysis technique in the serological diagnosis of dengue and Japanese encephalitis virus infections* Y. C. CHAN,' H. C. TAN, S. H. TAN, & K. BALACHANDRAN4 The single radial haemolysis testfor the serological diagnosis ofsuspected dengue and Japanese encephalitis virus infections uses crude virus antigens with a haemagglutinin titre of 1:320 or 1:640. The results, which may be read 3 hours after the addition ofa patient's serum, showed a general agreement between this test and haemagglutination-inhibition tests in the number of case diagnoses that were confirmed. The antibody responses of individual patients shown by the two tests, however, were different, which suggests that the two tests may not be measuring the same antibody. The single radial haemolysis test can dis- tinguish between dengue and Japanese encephalitis viruses using specific mouse hyper- immune sera. Tests on a limited number of sera from Japanese encephalitis patients also showed no cross-reactions with dengue virus antigens in those cases having a low-titred but significant fourfold antibody rise to Japanese encephalitis antigen. The single radial haemolysis (SRH) technique was first used for the measurement of influenza virus antibody (1, 2) and subsequently rubella antibody (3, 4). The technique has since been applied to other viruses including parainfluenza virus (5), mumps virus (6), coronavirus (7, 8), measles virus (9), reovirus (10), and bunyavirus (11). In addition to rubella virus, the SRH technique has been used with other togaviruses including dengue and Japanese encephalitis (JE) viruses (12-17). The SRH technique is based on radial diffusion of antibody in serum added to a well in an agarose gel containing antigen-sensitized sheep red blood cells (RBC) and complement, which may be added to the gel after diffusion of serum. Combination between antibody and antigen on the RBC in the presence of complement results in a circular zone of haemolysis, the diameter of which is proportional to the antibody concentration in the serum. The technique has been shown to be sensitive, accurate, and reproducible for the measurement of viral antibodies. This paper describes the application of the SRH technique to the * From the Department of Microbiology, Faculty of Medicine, National University of Singapore, Lower Kent Ridge Road, Singa- pore 0511, Singapore. Associate Professor. Requests for reprints should be addressed to this author. 2 Laboratory Technician. 3Laboratory Technician, Department of Pharmacy. 4Colombo Plan Trainee, Virus Research Institute, Department of Medical Sciences, Ministry of Public Health, Bangkok, Thailand. serological diagnosis of dengue and JE virus infections in patients and compares the results with those obtained by the commonly used haem- agglutination-inhibition (HI) test. A preliminary account of this work has been reported (18). MATERIALS AND METHODS Virus antigens The strains used were dengue virus type I (Hawaiian; DEN-1), dengue virus type 2 (New Guinea C; DEN-2), and JE virus (Nakayama). Crude antigens were used in both the HI and SRH tests, unless otherwise indicated. These antigens were prepared from 2007 suspensions of virus-infected suckling mouse brains in borate-saline buffer at pH 9.0 (BSB), which was centrifuged at 1500 g for 30 min. After being kept in a refrigerator (4 °C) for a few days, such crude virus antigens usually yielded haemagglutinin (HA) titres of 1:320 or 1:640 per 0.05 ml. Sucrose-acetone antigens were prepared according to the method of Clarke & Casals (19). Sera Sera were obtained from patients with a presumptive clinical diagnosis of dengue, dengue haemorrhagic fever, or Japanese encephalitis. Mouse hyperimmune sera were prepared by repeated 4616 -1043- 1044 Y. C. CHAN ET AL. intraperitoneal injections of adult mice with virus prepared from infected suckling mouse brains. Two sera from individual patients (304 and 319) and one serum pooled from several patients with high HI antibody titres (1: 1280 to 1: 5120) to DEN-I and DEN-2 antigens were used as positive control sera for the preparation of standard curves in the SRH test. Sera for the SRH test were heat-inactivated at 60 °C for 20 min in a sealed "U" disposable microtitration plate floating in a water bath. Sera for the HI test were treated with kaolin and absorbed with goose cells. SRH test Sheep RBC were washed three times in phosphate- buffered saline at pH 7.2 (PBS), and finally made up to a 10% suspension in borate-phosphate buffer (19) to obtain the optimal pH for haemagglutination of the virus (pH 6.2 for DEN-1, 6.4 for DEN-2, and 6.6 for JE). For sensitization, a 1007o RBC suspension was mixed with antigen in a ratio of 2: 1 and the mixture incubated at 4 °C for 10 min. The sensitized RBC were washed three times in PBS and finally resuspended in this buffer to make a 2.2% suspension. Stock solutions of 1.2% agarosea were prepared in PBS containing 0.1 7% sodium azide as a preservative. Two sources of complement were used: guinea pig serum stored frozen at - 70 °C and com- mercial lyophilized guinea pig complement.b To prepare SRH gels, appropriate volumes of sensitized RBC, agarose solution, and complement were mixed to give a final concentration of 1.007, 0.6%, and 4.0%7o, respectively. Three different types of plate were used to hold the gel: empty immunoplate (3 ml),' glass plate, 9.4 x 8.4 cm (12 ml), and square petri dish,d lO x 10 cm (15 ml). After the agarose had set, wells 2 or 3 mm in diameter were punched in the gel. In the test, a volume of 5 ll of serum was added to each well in the gel and the plate incubated in a moist chamber at 37 'C. The zones of haemolysis were measured after incubation for 3 h as well as overnight (18 h), using indirect illumination against a dark background. Limits of the zone diameter were first marked with a felt pen on the bottom of the plate and then measured with a transparent ruler to the nearest 0.5 mm. The SRH titre of a serum was determined graphically from the standard curve constructed from values of several dilutions of a positive control serum included in the test. For purposes of comparison with the HI titre, the SRH titre was converted to the nearest HI-equivalent titre. a Sigma Chemical Co., St. Louis, MO, USA. bWellcome Reagents Ltd, Beckenham, England. cHyland, Los Angeles, CA, USA. d Sterilin Ltd, Middlesex, England. HI test The procedure of Clarke & Casals (19) was followed with modifications for use in microtitration plates. In the test, a fourfold or greater dengue anti- body rise between acute- and convalescent-phase sera or a fixed antibody titre of at least 1: 1280 in both sera was considered to be significant evidence of a recent infection. RESULTS Since the main objective of this study was to develop the SRH technique as a simple and rapid test for antibody measurement, several variables which may affect its sensitivity and reproducibility were examined. Concentration ofsensitized RBC It was observed that as the concentration of sensitized RBC in the gel was reduced, so the diameter of the haemolytic zone (and thus the sensitivity of the test) increased. Using a 3-hour incubation time for diffusion of serum, no difference in the zone size was observed when red blood cells sensitized with DEN-2 virus were used in final concentrations of 0.507o, 1.0% and 2.0%. Reducing the RBC concentration to 0.25% produced larger zones (0.5-1.0 mm increase), but the contrast between lysed and unlysed areas of the gel was reduced. Concentrations of RBC of 1.0% and 2.0% gave the best contrast and the most distinct haemolytic zones and in the present study, a final 1.0% sensitized RBC concentration was used in all experiments. Concentration of virus The effect of the virus concentration used to sensi- tize the RBC on the diameter of the haemolytic zones was studied and the results are shown in Table 1. At low virus concentrations (HA titres of 1:40 and 1: 80), the haemolytic zones were indistinct and not measurable. The minimum virus concentration giving distinct and measurable zones was a HA titre of 1:160, and increasing the virus concentration to 1:640 HA titre did not significantly change the zone size. However, virus with HA titres of 1: 640 gave the most distinct haemolytic zones with all dilutions of the positive control serum while lower virus titres tended to produce zones with a fuzzy edge with higher serum dilutions. Sucrose-acetone extracted antigen The sucrose-acetone extracted antigen gave similar results as the crude antigen in the SRH test (Table 2). 1045SEROLOGICAL DIAGNOSIS OF DENGUE AND JAPANESE ENCEPHALITIS Table 1. Effect of virus concentration on the size of the haemolytic zones' Diameter of haemolytic zone (mm) for the following HA titres of virus used in sensitization of RBC: Serum dilutionb 640C 320 160 80 40 1:1 8.0 8.0 8.0 8.0 (NC)d 7.5 (NC) 1 :2 7.5 7.0 7.5 NCe 7.0 (NC) 1:4 6.5 6.5 7.0 NC NC 1:16 5.5 5.5 6.0 NC NC 1:64 4.5 4.5 5.0 (FE)f NC NC 0 DEN-2 gel incubated at 37 0C for 3 h. b Positive control serum with DEN-2 HI antibody titre of 1:1280. c Reciprocal of virus dilution. d NC = zone showing incomplete haemolysis but still measur- able. e NC = zone showing incomplete haemolysis and not measur- able. f Zone showing fuzzy edge. Table 2. Effect of different DEN-2 antigen preparations on size of haemolytic zones' Haemolytic zone diameter (mm) Crude antigen' Sucrose-acetone antigen d Serum dilution b (a) e (b) e Mean (a) (b) Mean 1:1 13.0 14.0 13.5 13.0 13.0 13.0 1:2 13.0 13.0 13.0 12.0 12.0 12.0 1:4 12.0 12.0 12.0 11.0 12.0 11.5 1:8 12.0 11.0 11.5 10.0 10.0 10.0 1:16 10.0 10.0 10.0 9.0 9.0 9.0 1:32 9.0 9.0 9.0 8.0 8.0 8.0 1:64 8.0 8.0 8.0 7.0 6.0 6.5 1:128 6.0 7.0 6.5 6.0 5.0 5.5 1:256 5.5 6.0 5.75 5.0 4.0 4.5 a Gel incubated at 37 OC for 18 h. b Dilutions of a positive control serum with DEN-2 HI antibody titre of 1: 2560. c 20% infected suckling mouse brain suspension centrifuged at 1500 g for 30 min. d Prepared by the method of Clarke & Casals (19). e Duplicate tests. However, the haemolytic zones produced by the crude antigen were larger than those obtained with the sucrose-acetone extracted antigen in all dilutions of the positive control serum, suggesting that the crude antigen may be more sensitive in the SRH test. Complement The effect of the source and concentration of complement, and the way in which complement was added to the gel on the development of haemolytic zones was examined. The activity of fresh guinea pig serum stored at - 70 °C and commercial lyophilized complement was compared (Table 3). Fresh guinea pig serum generally produced larger zones and a greater number of positive reactions. When lyophilized complement was used at a 1:30 dilution the zones were indistinct and not measurable. Complement may be incorporated into the gel before diffusion of serum, or added to the gel surface after the serum has been allowed to diffuse overnight at 4 'C. Of the two methods, larger haemolytic zones were obtained when the complement was incorpor- ated into the gel before diffusion of serum (Table 4). This method of adding complement to the gel was therefore used in this study. Table 3. Effect of complement source on the size of the haemolytic zones' Haemolytic zone diameter (mm) Fresh guinea Lyophilized Serum pig serumb complementc DEN-2 mouse hyperimmune serum diluted: 1:2 7.0 6.5 1:8 7.5 6.0 1:32 6.0 5.5 1:64 5.5 5.0 1:128 5.0 4.5 Patient 1 7.5 7.0 2 8.0 7.5 3 7.5 8.0 4 6.5 6.0 5 9.0 9.0 6 8.0 7.5 7 9.0 9.0 8 4.5 - 9 5.0 - 10 6.5 - 11 _d 12 - 13 - 0 DEN-2 gel incubated at 37 0C for 3 h. bPooled guinea pig serum stored frozen at - 70 OC. Wellcome product. d No haemolysis. Y. C. CHAN ET AL. Table 4. Effect of the method of adding complement on the size of the haemolytic zones Haemolytic zone diameter (mm) Complement incorporated Complement flooded in the gel before the over the gel after Serum addition of serum' serum diffusion b Positive control serum diluted: 1:1 14.0 13.0 1:2 13.0 11.5 1:4 11.5 10.5 1:16 9.0 8.0 1:64 6.5 (FE)' 5.0 Patient 1: acute _d convalescent 14.0 13.0 Patient 2: acute 9.5 10.0 convalescent 14.0 14.0 e DEN-2 gel incubated at 37 OC for 18 h after addition of serum. bDEN-2 gel incubated at 4 IC for 18 h after addition of serum and then at 37 OC for 3 h after addition of complement. c Zone showing fuzzy edge. d No haemolysis. Diffusion time In the SRH test, the serum is usually allowed to diffuse in the gel overnight (16-20 h) at 37 °C or at 4 °C followed by an additional incubation period at 37 °C for 2-3 h before the results are read. For a rapid test, the diffusion time would have to be shortened so that the result can be obtained on the day of the test. It was observed that haemolytic zones began to appear as early as one hour after the serum was added and incubated at 37 °C. Larger zones were obtained after overnight (18 h) incubation but there was a loss of clarity at the zone edge, and with some batches of RBC there was a complete loss of contrast between the areas of haemolysis and the background as a result of spontaneous haemolysis. Fig. 1 shows the relationship between serum dilu- tions and zone sizes of a positive control serum after incubation at 37 °C for 3 h and 18 h in a DEN-2 gel. In both cases, straight lines were obtained when the diameters of the zones were plotted against logio of the serum dilutions although the slope of the 3-h regression line was flatter. A similar linear relation- ship was obtained with DEN-1 virus, and the regres- sion lines of the two dengue serotypes were essentially 16 - *-* 3 hours o-o 18 hours DEN-2 14 F K U = I- U m U I- a s a MA =0 0 12 1 10- S 6 4 2 S 0 1:128 1:32 1:8 1:4 1:1 SERUM DILUTION Fig. 1. Linear relationship between haemolytic zone diameters and log1o of dilutions of a positive control serum following incubation at 37 OC for 3 h and 18 h. parallel (Fig. 2). This linear relationship was also observed with JE virus (unpublished data). When the regression lines produced by the same or different positive control sera in different batches of gel and tested on different days were compared, essentially parallel lines were obtained after 3 h and 18 h of incubation (Fig. 3 and 4). In the measurement of DEN-2 antibody in the patients' sera using diffusion times of 3 h and 18 h, no difference in sensitivity was observed (Table 5). Reproducibility Several dilutions of a positive control serum were tested from five to eight times on different days in different batches of gel and the combined results are shown in Table 6. The overall standard deviations for all serum dilutions were low, and the coefficient of variation for the 3-h test was about 6% and that for the 18-h test 5%. Differentiation between dengue and JE viruses When DEN-2 and JE mouse hyperimmune sera were tested against the two antigens by the SRH test, no cross-reactions between the two antigens were . . . . . 1046 SEROLOGICAL DIAGNOSIS OF DENGUE AND JAPANESE ENCEPHALITIS 9 _ 8 E E IC 7 w IC a 6 z 0 N 2 5 I- 0 U<AM4 3 *-4 DEN-1 0-0 DEN-2 3 hours 0y S 6 42I- S MO a us 0 Li a 4 2 1:256 1:64 1:16 1:4 1:1 SERUM DILUTION Fig. 2. Regression lines of DEN-1 and DEN-2 viruses following incubation at 37 OC for 3 h. The lines are almost parallel. Table 5. Effect of diffusion time of serum on the sensi- tivity of the SRH test DEN-2 antibody titre measured by SRH after serum diffusion for: Serum Hi titre 3 hours 18 hours Patient 1 < 10" < job < 10 2 < 10 < 10 < 10 3 20 80 40 4 40 80 80 5 80 160 80 6 80 160 80 7 1280 2560 1280 8 (a)' < 10 < 10 < 10 (c)d 160 640 1280 9(a) 160 160 80 (c) 1280 1280 1280 a Reciprocal of serum dilution. b Reciprocal of serum dilution. The SRH titre is the nearest HI- equivalent titre determined from the standard curve. ((a) = acute-phase serum. d (c) = convalescent-phase serum. 0 1:256 1:64 1:16 1:4 1:1 1:256 1:64 1:16 1:4 1:1 SERUM DILUTION Fig. 3. Regression lines of DEN-2 virus obtained with the same positive control serum in four different tests incubated at 37 OC for 3 h. The lines are almost parallel. 14O a a s*0 SER*U:M4I:1 t:4 t:l SERIUM DIUTION Fig. 4. Regression lines of DEN-2 virus obtained with two different positive control sera (No. 304 and 319) in different tests incubated at 37 OC for 18 h. The lines are almost parallel. 1047 ..... . .... . .... . ... ... .. . .... . .... . . ... .. . .. . ... .. ...I 319 311 Y. C. CHAN ET AL. Table 6. Variation in size of haemolytic zones using different serum diffusion times and in gels from different batches' Haemolytic zone diameter (mm) Dilution of positive control serum 3 hours 18 hours 1:1 8.11 ±0.22b 14.33±0.53c 1:2 7.36±0.24 13.00±0.00 1:4 6.88±0.44 12.25±1.04 1:16 6.06±0.50 9.58±0.92 1:64 4.93±0.53 ° Results pooled from different tests using DEN-2 virus and the same positive control serum. b Mean ± 1 standard deviation. The mean for each serum dilution was calculated from diameters obtained from 7-8 tests. The overall coefficient of variation for all serum dilutions was 6.27%. ' Mean ± 1 standard deviation. The mean for each serum dilu- tion was calculated from diameters obtained from 5-6 tests. The overall coefficient of variation for all serum dilutions was 5.43%. observed (Table 7). This specificity of the SRH test in the differentiation of dengue and JE viruses was again demonstrated in an experiment comparing HI and SRH antibody titres to DEN-1, DEN-2, and JE antigens (Table 8). Unlike the HI test which showed cross-reactivity among the three antigens, the JE antiserum reacted only with its own antigen. Serological diagnosis ofdengue virus infection Paired sera from 21 patients with a presumptive diagnosis of dengue or dengue haemorrhagic fever were tested for DEN-I and DEN-2 antibodies by both HI and SRH tests and the results are shown in Table 9. By the HI test, 10 out of 21 patients were confirmed as having a recent dengue virus infection, while 9 out of 21 patients were confirmed by the SRH test. There were two patients (No. 3 and 8) who were confirmed Table 7. Specificity of the SRH test in the differentiation of dengue and Japanese encephalitis viruses' Haemolytic zone diameter (mm) Dilution of mouse hyperimmune serum DEN-2 virus JE virus DEN-2 1:2 7.0 - 1:8 6.0 - 1:32 5.0 - 1:64 4.0 - 1:128 _b _ JE 1:2 - 11.0 1:8 - 9.0 1:32 - 6.0 1:64 - 5.0 1:128 - 4.0 ° Gels were incubated at 37 OC for 3 h. b No haemolysis. by the HI but not by the SRH test. These 2 patients had very low HI antibody rises, from < 1: 10 in the acute serum to 1: 20 in the convalescent serum. On the other hand, 2 patients (No. 9 and 12) who were not confirmed by the HI test had significant antibody rises by the SRH test. When the HI and SRH antibody titres in individual patients were compared, no agreement was seen in many cases, especially those sera containing no dengue antibody (titre < 1:10) and those showing a significant fourfold antibody conversion. The overall findings may be summarized as follows. (1) In the HI test, paired sera from 2 patients were negative (titre < 1: 10) to one or both dengue antigens whereas there were 15 such patients by the SRH test. (2) In the HI test, there were 6 patients who had significant anti- body rises to both dengue antigens, and 4 showing rises to either DEN-1 or DEN-2. In contrast, no Table 8. Comparison of HI and SRH tests in the differentiation of dengue and Japanese encephalitis viruses' HI test with virus SRH test with virus Mouse hyperimmune serum DEN-1 DEN-2 JE DEN-1 DEN-2 JE DEN-1 320b 80 160 320' 40 < 10 DEN-2 40 160 320 NCd 160 < 10 JE 80 160 1280 <10 <10 1280 a Gels were incubated at 37 OC for 3 h. b Reciprocal of serum dilution. ' Reciprocal of serum dilution. The SRH titre is the nearest HI-equivalent titre determined from the standard curve. d Zone showing incomplete haemolysis and not measurable. 1048 SEROLOGICAL DIAGNOSIS OF DENGUE AND JAPANESE ENCEPHALITIS Table 9. Serological diagnosis of dengue virus infection by Hi and SRH testsa Hi titre against SRH titre against Patient's serum Date of serum DEN-1 DEN-2 DEN-1 DEN-2 26-06-82 16-07-82 06-07-82 21-07-82 06-07-82 21-07-82 19-07-82 23-07-82 05-08-82 27-07-82 02-08-82 19-08-82 01-09-82 20-08-82 24-08-82 21-08-82 13-09-82 0 26-08-82 08-09-82 1 02-09-82 15-09-82 2 11-10-82 19-10-82 3 29-10-82 11-11-82 4 01-11-82 11-11-82 5 10-11-82 22-11-82 6 10-11-82 22-11-82 17 16-11-82 19-11-82 18 17-11-82 27-01-83 19 11-03-83 30-04-83 20 28-04-83 10-05-83 21 31-05-83 11-06-83 0 SRH gels were incubated at 37 OC for 3 h. bReciprocal of serum dilution. 20 < 10 < 10 < 10 160 < 10 40 20 20 320 320 20 80 20 20 640 640 < 10 20 320 320 20 40 < 10 20 640 640 40 40 20 20 20 10 < 10 < 10 40 40 20 20 160 2560 640 640 < 10 20 20 < 10 20 < 10 20 < 10 20 < 10 20 10 10 640 640 < 10 < 10 320 320 < 10 20 < 10 20 640 640 20 20 20 20 10 10 < 10 < 10 40 40 20 20 20 160 640 640 < 10 20 < 10 < 10 40 < 10 < 10 < 10 40 < 10 320 < 10 < 10 320 320 < 10 < 10 < 10 40 < 10 80 < 10 < 10 1280 640 160 160 < 10 < 10 < 10 < 10 < 10 < 10 80 40 < 10 < 10 640 320 640 320 < 10 < 10 <10 <10 <10 <10 <10 <10 <10 <10 <10 20 20 320 320 <10 <10 80 80 <10 <10 <10 80 320 1280 20 20 20 40 <10 <10 <10 <10 80 80 <10 <10 40 160 1280 640 <10 20 1 1049 I Y. C. CHAN ET AL. Table 10. Serological diagnosis of Japanese encephalitis virus infection by CF, Hi, and SRH tests' Hi titre against: SRH titre against: Serum CF titre against JE JE DEN-2 JE DEN-2 Patient 1 (a)b < 8c 160' 160 < 10c < 10 (c)b 64 NDd ND 80 < 10 Patient 2 (a) <8 ,40 < 10 80 < 10 (c) 32±e NSf NS 320 < 10 Patient 3 (a) 32 ± 160 80 160 20 (c) )128 5120 1280 2560 160 Patient 4 (a) <8 20 < 10 < 10 < 10 (c) 8± >40 10 160 <10 Patient 5 (a) <8 ND ND <10 <10 (c) <8 ND ND < 10 < 10 a SRH gels were incubated at 37 IC for 3 h. b (a) = acute-phase serum; (c) = convalescent-phase serum. c Reciprocal of serum dilution. d ND = not done. ± ± indicates less than maximal complement fixation at the indicated serum dilution. f NS = serum showing non-specific haemagglutination. patient had a significant antibody rise to both antigens in the SRH test; however, there were 9 patients who showed a significant antibody rise to one dengue antigen, 5 to DEN-1, and 4 to DEN-2. (3) HI antibody rises were generally of low titres, from < 1: 10titre in the acute-phase serum to a titre of 1: 20 or 1:40 in the convalescent-phase serum. On the other hand, these sera in the SRH test showed antibody titres generally above 1:40. (4) In those patients showing no dengue antibody rise but fixed antibody titres to the two antigens in their paired sera in both HI and SRH tests, the titres obtained in the two tests were comparable, particularly in those sera with HI titres of 1: 40 and above. Serological diagnosis ofJE virus infection Acute- and convalescent-phase sera from patients with clinically diagnosed Japanese encephalitis were tested by the complement fixation (CF), HI, and SRH tests. The CF test is known to be more specific than the HI, and is generally capable of differentiating Japanese encephalitis from dengue virus infection, particularly in a primary-type of infection (i.e., the acute serum has no detectable JE antibody while the convalescent serum shows a low-titred but significant antibody rise). In contrast, the HI test, in most cases, shows antibody rises to both JE and dengue antigens. The results showed that the SRH test closely resembles complement fixation in specificity in distinguishing between the two virus infections (Table 10). DISCUSSION The main purpose of this study was to evaluate the SRH technique as a simple and rapid test for the sero- logical diagnosis of dengue and JE virus infections. It has been suggested that the SRH test may be a useful substitute for the HI test for the measurement of anti- bodies to rubella and influenza viruses. The SRH test is simpler to perform than the HI test, as the former does not require serial dilutions of serum and pre- treatment of serum to remove non-specific inhibitors and naturally occurring agglutinins for RBC. The only treatment of serum needed for the SRH test is heating at 60 'C for 20 min. One of the main criticisms of the SRH test was that 1050 SEROLOGICAL DIAGNOSIS OF DENGUE AND JAPANESE ENCEPHALITIS it requires a large amount of high-titred antigen for sensitization of RBC (20). The present study showed that, although the SRH test consumed more antigen than the HI, the titre oT the antigen need not be high. Dengue virus antigen with HA titres as low as 1: 160 per 0.05 ml was found to be adequate for sensitiz- ation of RBC. Furthermore, crude virus antigens, which generally yielded a HA titre of 1: 320 or 1: 640, could be used in the SRH test, thus eliminating the need for the elaborately produced sucrose-acetone extracted antigen recommended for the HI test. The SRH test, unlike the HI, is also unaffected by dif- ferences in antigen concentration, although clearer and sharper haemolytic zones were produced when higher antigen concentrations above a minimal amount were used. This relative insensitivity of the SRH test to large changes of antigen concentration has also been observed in other viruses (1, 2, 4, 5, 9, 12, 16, 17, 21). Although the concentration of complement did not affect the size of the haemolytic zone produced but only its clarity, it was found that the source of com- plement was important. Fresh guinea pig serum appeared to produce larger zones and is a better source of complement than the lyophilized product, which was also noted by Grilli & Davies (22). In the present study, the haemolytic zone diameter was measured after serum had been diffused for 3 h at 37 °C instead of the usual overnight incubation. A 3-h incubation time was selected in order to provide the clinician with the test result on the day of submis- sion of the blood specimen to the laboratory. It was found that shortening of the diffusion time to 3 h affected neither the sensitivity nor the reproducibility of the SRH test. Vaananen (17), working with Semliki Forest virus, also found that the SRH reaction could be read after two hours of incubation. Grilli & Davies (22) found that the optimal incubation time for influenza B virus was six hours at 37 'C. Most workers employing the SRH technique related the size of the haemolytic zones to the HI titres since the SRH measurements were not directly inter- pretable. Although the scatter diagrams of the haemolytic zone diameter against the HI titre generally showed a good correlation in many studies, discrepancies between the SRH and HI tests have also been reported (1, 2, 8, 10, 21, 23-28). It is not known whether the antibody measured by the SRH test is the same as that measured by the HI. Klingeborn & Dinter (29) found that the equid herpesvirus antibody mediating the SRH reaction was probably the one responsible for neutralization. Champsaur et al. (30) also observed that rubella SRH antibodies, like the CF and neutralizing antibodies, rose later than the HI antibodies during the course of a primary infection. In the present study, there was general agreement between the HI and SRH tests in the total number of patients who were confirmed to have had a recent dengue infection. However, differences in antibody response to the two dengue antigens (DEN-1 and DEN-2) in individual patients were observed. By the HI test, 6 out of 10 of the confirmed patients had antibody rises to both dengue antigens while the remainder had rises to one of the two antigens. In contrast, by the SRH test, none of the patients showed an antibody rise to two dengue antigens and 7 out of 10 confirmed patients had an antibody rise to only one antigen. Since virus was not isolated from the patients and only two of the four dengue serotypes were used in the tests, it is not possible to make a definitive etiological diagnosis in any given case or to conclude on the serotype specificity of the SRH test. The specificity of the SRH test in differentiating members of the alphaviruses and flaviviruses has been reported (12-14, 17). Duca et al. (13) and Gaidamovich & Melnikova (14) found no cross- reactions between dengue and West Nile viruses in the SRH test. No cross-reactions between dengue and JE viruses were observed in the present study. From the limited data presented, it also appeared that the SRH test can differentiate between dengue and JE infections in patients. A similar specificity of the SRH test was suggested in the findings of Gaidamovich et al. (15) from a serological survey for dengue and JE antibodies among febrile patients in Bangladesh. With specific mouse hyperimmune sera, JE virus con- sistently did not show cross-reactions with the four dengue virus serotypes (unpublished data). The SRH test therefore shows great promise for sero-epidemiological studies and laboratory diag- nosis in areas where both dengue and JE viruses are endemic. ACKNOWLEDGEMENTS This work was supported by a research grant from the Shaw Foundation, Singapore. The JE patients' sera and CF test results were kindly provided by Dr S. Doraisingham, Department of Pathology, Ministry of Health, Singapore. 1051 1052 Y. C. CHAN ET AL. RESUME EMPLOI D'UNE TECHNIQUE D'HEMOLYSE RADIALE SIMPLE POUR LE DIAGNOSTIC SEROLOGIQUE DE LA DENGUE ET DES INFECTIONS DUES AU VIRUS DE L'ENCEPHALITE JAPONAISE Le principal objectif de cette etude etait de developper une technique d'hemolyse radiale simple (SRH) pour disposer d'une epreuve simple et rapide pour le diagnostic s6rolo- gique de la dengue et des infections dues au virus de l'enc& phalite japonaise. Divers parametres qui peuvent affecter la sensibilite et la reproductibilite de l'epreuve ont e examines. II a e observ6 que la concentration des hematies sensibilis6es dans le gel d'agarose modifiait le diametre de la zone d'hemolyse obtenue. Avec un temps d'incubation de 3 h, aucune difference dans les dimensions de la zone n'a e observ&e avec des hematies sensibilis&es par le virus de la dengue et utilis6es a une concentration finale de 0,5% ou plus. Un virus de la dengue ayant un titre d'hemagglutinine (HA) de 1/160 pour 0,05 ml suffisait a sensibiliser les hematies, et l'augmentation de la concentration de virus ne modifiait pas la dimension de la zone. Les antigEnes viraux bruts, pr6par6s par centrifugation a faible vitesse de suspensions de cerveaux infect6s de souriceaux a la mamelle, peuvent etre utilises pour l'epreuve SRH. Ces antigenes ont generalement un titre d'hemagglutinine de la dengue de 1/320 ou de 1/640. Le serum frais de cobaye s'est montre superieur au complement Iyophilise pour l'execution de 1'epreuve. Une dur&e d'incubation de 3 h a 37°C a e compar6e avec l'incubation habituelle de la nuit entiere et il a e constate qu'elle n'affecte pas la sensibilite ni la reproductibilite de l'6preuve. Des droites ont e obtenues quand les diametres des zones d'hemolyse ont e rapportes au log 10 des dilutions d'un serum temoin positif contenant un titre eeve d'anticorps, mesure par inhibition de l'hemag- glutination (HI), envers la dengue ou le virus de l'enc6pha- lite japonaise. Dans la totalite des cas positifs, les resultats des epreuves faites sur des serums de malades atteints de dengue ou de dengue hemorragique montraient dans l'ensemble une bonne correlation entre les epreuves HI et SRH. Cependant, chez certains malades, la reponse en anticorps au virus de la dengue des types 1 et 2 montrait des differences faisant penser que les deux epreuves pouvaient ne pas mesurer le meme anticorps. Avec des serums specifiques de souris hyperimmunes, 1'epreuve SRH n'a pas montre de reactions croisees entre les virus de la dengue et de l'encephalite japonaise. La SRH r6alisee sur un petit nombre de malades atteints d'encepha- lite japonaise a paru montrer une specificite analogue, sans reactions croisees avec les antigenes du virus de la dengue dans les serums oii une primo-infection par le virus de l'en- cephalite japonaise pouvait etre mise en evidence. REFERENCES 1. SCHILD, G. C. ET AL. Single-radial-haemolysis: a new method for the assay of antibody to influenza haemag- glutinin. Bulletin of the World Health Organization, 52: 43-50 (1975). 2. RUSSELL, S. M. ET AL. A single radial haemolysis tech- nique for measurement of influenza antibody. Journal of general virology, 27: 1-10 (1975). 3. SKAUG, K. ET AL. Application of the passive haemolysis test for the determination of rubella virus antibodies. Acta pathologica et microbiologica Scandinavica, Section B, 83: 367-372 (1975). 4. STRANNEGARD, 0. ET AL. Hemolysis-in-gel test for the demonstration of antibodies to rubella virus. Journal of clinical microbiology, 1: 491-494 (1975). 5. PROBERT, M. & RUSSELL, S. M. Measurement of para- influenza-3 virus antibody by the single radial hemolysis technique. Journal of clinical microbiology, 2: 157-161 (1975). 6. VAANANEN, P. ET AL. Determination of mumps and influenza antibodies by haemolysis-in-gel. Archives of virology, 52: 91-99 (1976). 7. HIERHOLZER, J. G. & TANNOCK, G. A. Quantitation of antibody to non-hemagglutinating viruses by single radial hemolysis: serological test for human corona- virus. Journal of clinical microbiology, 5: 613-620 (1977). 8. RISKI, H. ET AL. Antibodies to human coronavirus OC43 measured by radial haemolysis in gel. Scandina- vian journal of infectious diseases, 9: 75-77 (1977). 9. WESSLEN, L. Demonstration of antibodies to measles virus by hemolysis-in-gel (HIG) test. Scandinavian journal of infectious diseases, 10: 15-20 (1978). 10. D'AMBROSIO, E. ET AL. Single radial haemolysis for the determination of antibody to reoviruses. Archives of virology, 68: 309-312. 11. MALKOVA, D. ET AL. Radial haemolysis in gel for detection of antibodies to bunyavirus Lednice (Turlock group). Acta virologica, 27: 439-441 (1983). 12. ODELOLA, H. A. Application of single-radial- haemolysis test for the detection of antibodies to toga- viruses. Archives of virology, 60: 325-328 (1979). 13. DUCA, M. ET AL. Single radial haemolysis for the assay of antibodies to some haemagglutinating arboviruses. Bulletin ofthe World Health Organization, 57: 937-942 (1979). 14. GAIDAMOVICH, S. Y. & MELNIKOVA, E. E. Passive hemolysis-in-gel with Togaviridae arboviruses. Inter- virology, 13: 16-20 (1980). 15. GAIDAMOVICH, S. Y. ET AL. Serological evidence of dengue fever in the Bangladesh Republic. Acta viro- logica, 24: 153 (1980). SEROLOGICAL DIAGNOSIS OF DENGUE AND JAPANESE ENCEPHALITIS 1053 16. GAIDAMOVICH, S. Y. ET AL. Radial haemolysis in gel used in serological studies on Venezuelan equine ence- phalomyelitis virus. Acta virologica, 25: 36-40 (1980). 17. VAANANEN, P. The use of red cells with fused Semliki Forest virus envelope proteins in antibody determi- nations by hemolysis in gel. Journal of virological methods, 4: 117-126 (1982). 18. CHAN, Y. C. ET AL. Single radial haemolysis test for the routine diagnosis of dengue virus infections. In: Pro- ceedings of the International Conference on Dengue/ Dengue Haemorrhagic Fever, Kuala Lumpur, 1-3 September 1983. Kuala Lumpur, 1983, pp. 466-477. 19. CLARKE, D. & CASALS, J. Technique for hemag- glutination and hemagglutination-inhibition with arthropod-born viruses. American journal of tropical medicine & hygiene, 7: 561-573 (1958). 20. WHO Technical Report Series, No. 661, 1981 (Rapid laboratory techniques for the diagnosis of viral infections: report of a WHO Scientific Group), pp. 13-15. 21. FORGER, J. M. & GILFILLAN, R. F. Single radial hemolysis as a cost-effective determinant of rubella antibody status. Journal of clinical microbiology, 9: 115-119 (1979). 22. GRILLI, E. A. & DAVIES, J. R. Influenza B at Christ's Hospital: natural antibody to influenza B estimated by radial haemolysis. Journal ofhygiene (Cambridge), 87: 211-218 (1981). 23. CLARKE, M. ET AL. The use of single-radial-haemolysis for rubella antibody studies. Journal of hygiene (Cambridge), 79: 355-364 (1977). 24. OXFORD, J. S. ET AL. Quantitation and analysis of the specificity of post-immunization antibodies to influenza B viruses using single radial haemolysis. Journal of hygiene (Cambridge), 88: 325-333 (1982). 25. HARNETT, G. B. ET AL. Single-radial hemolysis test for the assay of rubella antibody in antenatal, vaccinated, and rubella virus-infected patients. Journal of infectious diseases, 140: 937-944 (1979). 26. PETTERSSON, R. F. Correlation between antibody titres to influenza A virus subtypes HINI and H3N2 as measured by haemagglutination-inhibition and haemolysis-in-gel tests. Journal of infectious diseases, 142: 450 (1980). 27. MANCINI, G. ET AL. Comparison of haemagglutination- inhibition and single radial haemolysis techniques for detecting antibodies to influenza A and B viruses. Journal of hygiene (Cambridge), 91: 157-162 (1983). 28. KURTZ, J. B. ET AL. Rubella antibody measured by radial haemolysis. Characteristics and performance of a simple screening method for use in diagnostic laboratories. Journal of hygiene (Cambridge), 84: 213-222 (1980). 29. KLINGEBORN, B. & DINTER, Z. Measurement of neutralizing antibody to equid herpesvirus I by single radial hemolysis. Journal of clinical microbiology, 7: 495-496 (1978). 30. CHAMPSAUR, H. ET AL. Hemagglutination inhibition, single radial hemolysis, and ELISA tests for the detection of IgG and IgM to rubella virus. Journal of medical virology, 5: 273-286 (1980).