Journal of Infection ([99I), zz, t7-26

Diagnostic value of clinical and bacteriological findings in pertussis Gunnar Granstr6m,*ll Bengt Wretlindt and Marta Granstr6m~ * Department of Infectious Diseases, ~ Department of Bacteriology, Danderyd Hospital, S - I 8 2 88 Danderyd, Sweden, ~ Department of Vaccine Production, National Bacteriological Laboratory, S- Io5 21 Stockholm and t Department of Clinical Microbiology, Karolinska Hospital, Box 60 5oo, S-IO4 oI Stockholm, Sweden Accepted for publication I I June I99O

Summary Clinical and bacteriological findings in the diagnosis of pertussis were evaluated in 3oo consecutive patients with parental or the patient's own suspicion of the disease. Serology was used as a reference method. Of the 285 (95 %) patients fully sampled, I63 (57%) were diagnosed as having pertussis while the remaining Iz2 patients constituted the non-pertussis control group. The clinical and epidemiological data were collected at the first visit made on median day seven of illness. In this population of mainly unimmunised children, the highest predictive values were obtained for the physician's diagnosis of pertussis (Ioo %) and for the physician's diagnosis of some other illness (93 %). The only clinical symptom with a high predictive value for pertussis was the report of whoops (92 %). Among epidemiological data the highest predictive value (90 %) was obtained for reported household exposure in unimmunised children more than I year of age. Culture of Bordetella pertussis was found to have an overall 5° % sensitivity. Isolation of other bacteria had no predictive value in the differential diagnosis of pertussis.

Introduction Both clinical and laboratory diagnosis of pertussis are reputedly difficult. T h e most reliable, i.e. specific, m e t h o d is the culture of Bordetella pertussis, b u t it has the drawback of low sensitivity. 1'2 Evaluation of other diagnostic m e t h o d s with culture as a reference m e t h o d entails the risk that the control group of non-pertussis patients will include m a n y culture-negative cases of pertussis. Recent developments in pertussis serology have greatly i m p r o v e d the sensitivity of laboratory diagnosis. 3'4 Serological tests m a y therefore be useful as a reference in the evaluation of other diagnostic methods. An enzyme-linked i m m u n o s o r b e n t assay ( E L I S A ) with filamentous haemagglutinin ( F H A ) and pertussis toxin ( P T ) as antigens was recently f o u n d to have Ioo ~o sensitivity w h e n evaluated in 90 culture-confirmed cases of pertussis. 5 Its sensitivity in a culture-negative population m a y be lower, b u t should be sufficient to allow the differentiation of pertussis from other respiratory infections with an error of little significance in a large population of patients. Although serological testing is diagnostically sensitive, it has the disI] Address correspondence to: Dr Gunnar Granstr6m, Department of Infectious Diseases, Danderyd Hospital, 8-I82 88 Danderyd, Sweden. oi63-4453/9I/OiOOI7+ to $03.00/0

© I99I The British Society for the Study of Infection

18

G. GRANSTR(3M E T A L .

advantage of being slow and may not be generally available. Culture for B. pertussis is also a relatively slow method, a n d failure to isolate the bacterium does not exclude the disease. T h e most widely used and also the most rapid diagnostic m e t h o d is the evaluation made by a physician based on clinical symptoms and epidemiological data. Clinically diagnosed cases also represent the basis of the notification systems used for the surveillance of pertussis in most countries. T h e development of serological methods has for the first time given a reference m e t h o d which allows a scientific evaluation of the diagnostic sensitivity, specificity and predictive value of clinical symptoms, epidemiological data and physicians' evaluation in the diagnosis of pertussis. T h e aim of the present study was to investigate these features as well as the sensitivity, specificity and predictive value of microbial findings in patients with respiratory symptoms. Materials and methods Patients

Three h u n d r e d consecutive patients with suspected pertussis entered the study between August 1986 and May 1987 . All but three came with parental or the patient's own suspicion of the disease to the out-patient clinic of the D e p a r t m e n t of Infectious Diseases, Danderyd Hospital. T h r e e patients were referred from other clinics for admission to hospital with suspected pertussis. More than 9o % patients were seen at the first visit by the same physician (G. G.). T h e y received an appointment for a second visit 6 weeks later. Of the 30o patients, eight were excluded because of technical errors in sampling and seven for refusal of a second blood sample. T w o further children refused a second blood sample but could be included on the strength of a positive culture for B. pertussis. T h e 285 (95 %) patients were diagnosed by laboratory methods (culture and serology) to have either pertussis (n = 163) or some other disease (n = 122). T h e pertussis patients had a median age of 3"7 years (range o.3-63-2 years). One h u n d r e d and fifty (92 %) were children (less than 15 years of age) and included 13 (8 %) infants under I year of age. Eight children, median age lO"3 years (range 8"o-14"6 years) and born before general immunisation ended, had received three doses of diphtheria-tetanus-pertussis ( D T P ) vaccine as infants. T h e remaining 142 children were u n i m m u n i s e d . Of the 13 adults, four reported D T P - i m m u n i s a t i o n as infants. Eighty (49%) of the pertussis patients were females. T h e 122 non-pertussis patients had a median age of 2"3 years (range 0"2-49"8 years). Of these, lO7 (88 %) were children among w h o m 28 (23 %) were infants. Seven of the children (median age 8"4 years, range 4"8-13"3 years) had received three doses of D T P vaccine as infants. T h e remaining ioo children were unimmunised. T h r e e of the 15 adults reported immunisation with D T P vaccine as infants. Sixty-one (5o%) of the nonpertussis patients were females. For the pertussis patients, the median duration of illness was 7 days (range o-96 days) at the time of the first visit. T h e corresponding median for the nonpertussis patients was 6 days (range o-151 days).

Diagnosis of pertussis

I9

T h e study was approved by the local ethical committee and sampling was performed after informed consent had been obtained from the patients or their parents.

Sampling and handling of samples At the first visit, a sample of nasopharyngeal secretion was obtained from each patient by suction through fine plastic tubing introduced into the nasopharynx. 6 T h e material aspirated was cultured for B. pertussis and other bacteria within 2 h of sampling. A venous or capillary blood sample was drawn from adults and children respectively at both the first and a second visits. T h e duration of illness at the time of the first blood sample was the same as that at the first visit while the second sample was drawn on median day 43 (range 34-I25 days) after the first sample from the pertussis patients and on median day 44 (range 3o-I75 days) from the non-pertussis patients. Each serum sample was stored at - 7 0 °C until the second sample was obtained, when both samples were tested in parallel by E L I S A and by the neutralisation test (NT).

Bacterial culture T h e aspirates were cultured for B. pertussis on Bordet-Gengou agar for 5 days. T h e Bordet-Gengou agar contained 17 % defibrinated horse-blood and 2 mg cloxacillin. It was freshly prepared each week. Isolates of B. pertussis were confirmed by agglutination and by immunofluoresence. Culture for other bacteria was carried out on blood agar and haematin agar plates incubated for 1 day. No attempt was made to measure the growth, neither was every colony identified. T h e cultures were read mainly be the same physician (G.G.), applying the same criteria. In the comparative study of media for isolating B. pertussis, the nasopharyngeal swabs were also streaked on charcoal agar with Io % defibrinated horse-blood and 4o mg cephalexin. T h e medium was freshly prepared each week. 7

Serology and serological criteria E L I S A was done with pertussis toxin (gift of the Osaka Foundation, Biken, Japan) and filamentous haemagglutinin (preparation of the National Bacteriological Laboratory, Stockholm, Sweden) as previously described in detail. 3'~ T h e N T was performed in Chinese hamster ovary (CHO) cells after the ELISA. s'9 A patient was defined as serologically positive for pertussis if either (z) a significant (at least two-fold) rise in titre between the two samples was demonstrated either in IgG to PT, or in IgG or IgA to FHA, or at least a four-fold rise by N T or (2) a corresponding significant decrease in the same test and antibody classes were found or (3) high unchanging titres of IgA to F H A were found in both samples, s' ~ T h e 80 patients with positive cultures for B. pertussis and two blood samples were all positive by criterion (Q. In addition, 65 patients with negative cultures for B. pertussis were positive by this criterion. T h e remaining I6 patients were positive by criterion (2) (eight cases, including four adults) or by criterion (3) (eight cases, including three adults).

20

G. GRANSTRt3M

ET AL.

Clinical diagnosis

T h e physician's diagnosis of pertussis was made in connection with the physical examination at time of the first visit. It was based on noted and reported clinical features as well as on epidemiological data, but without results of laboratory tests. Statistical analysis and calculations

T h e data were analysed on a personal computer by means of the data base program dBase I I I Plus (Ashton-Tate). Differences between groups were compared by Fisher's exact two-tailed test. Results Clinical findings and evaluations

T h e major clinical features in the 285 patients at the time of their first visit are shown in Table I. All patients had a cough. Vomiting related to coughing was reported to a similar extent in the two groups. Fever was u n c o m m o n , but more frequent in the non-pertussis patients. Of the seven pertussis patients with fever, two had a concomitant varicella infection and one entered the study when admitted to hospital with pneumonia. Of the non-pertussis patients with fever, two had pneumonia and two had otitis media. Reporting of whoops was significantly more c o m m o n for patients with pertussis. T h e physician ventured a clinical diagnosis in Ioo/285 (35 %) patients (Table I). T h e distribution showed that the disease is more easily diagnosed than excluded (45 % v s . 20 %, P < o.ooI). T h e clinical diagnosis of pertussis was confirmed by laboratory methods in all cases. T h e clinical diagnosis of non-pertussis was erroneous in two cases, an infant of 4 m o n t h s and a I5-yearold girl who had been i m m u n i s e d in infancy with three doses of D T P vaccine. Information on the duration of cough at the time of the second sample was obtained from 202/285 (71%) patients. A cough of 6 weeks or more was reported by 3I % (39/I27) patients with pertussis and by 27 % (20/75) with non-pertussis disease (no significant difference). Four of I5O children with pertussis were admitted to hospital but three entered the study because of admission to hospital. Of the I47 children entering the study as out-patients, a total of 17 (I2 %) suffered a complication. One 2.2-year-old child was admitted to hospital for suspected pertussis encephalopathy. Six children had pneumonia, eight had otitis media, one sinusitis and one received antibiotic therapy for fever. None of the I3 adults with pertussis had any complications. A m o n g the non-pertussis patients, none was admitted to hospital. Six (5"6 %) of Io7 children presented with or later developed p n e u m o n i a 2 or otitis media. 4 Of the I5 non-pertussis adults, one had p n e u m o n i a and one otitis media at the first visit. T r e a t m e n t with erythromycin for pertussis was given to 82/I63 (5o%) patients with pertussis. Of the children who later developed complications of pneumonia and otitis media, 9 / I 4 (64%) received this treatment. Before entering the study, 5/163 (3 %) patients had received erythromycin. In two of these cases, erythromycin was given as prophylaxis but the children still developed clinical and laboratory-confirmed disease. One culture-positive

Diagnosis of pertussis

2I

Table I Clinical symptoms and physicians' evaluations in patients with

pertussis and pertussis-like disease at the first visit Pertussis*

Nonpertussis Number of patients with symptoms (%) n = I63 n = I22 P-value

Cough with vomiting Whoops Fever /> 38 °C Physicians' diagnosis Pertussis Non-pertussis

28 (I7) 60 (37) 7 (4)

I6 (I3)

5 (4) 15 (I2)

N.S. < O'OOI o-o23

73 (45) 2 (I)

o (o) 25 (20)

< o-ool < o'ooi

* Two cases of parapertussis included. N.s. = N o t s i g n i f i c a n t .

child had had culture-confirmed pertussis 3 years previously. E r y t h r o m y c i n was given during the first episode but was declined in the present study. E r y t h r o m y c i n for suspected pertussis was given to 3 5 / I 2 2 ( 2 9 % ) nonpertussis patients. N o n e of the four patients who later developed complications had received this treatment.

Epidemiological data T h e types of reported exposure to pertussis are shown in Table II. T h e most c o m m o n exposure was in day-care, reported for I33/285 (47%) patients. A m o n g these patients, the disease was laboratory-confirmed in 73 % (89/I22) u n i m m u n i s e d children I - 9 years of age. T h i s rate is not significantly different from the 9o % (I8/2o) for household exposure in the same group. A similar rate of 8 2 % ( 9 / I I ) laboratory-confirmed cases was found in the group of adults and older children who had received three doses of D T P vaccine as infants. T h e 42 % (5/12) rate for laboratory-confirmed pertussis in the infants was, however, significantly lower than in the older u n i m m u n i s e d children (P = o'oi2). T h e distribution of laboratory-confirmed cases u p o n household exposure showed 2/7 positives for infants 0-6 m o n t h s of age and 3/5 for infants 7 - I 2 m o n t h s of age. O f the 28 infants with negative laboratory findings, seven had a sibling with laboratory-confirmed disease within the study, including the two cases of parapertussis. In a further five cases, the disease was excluded also in the alleged contact. Exposure at school was reported for nine children, with seven (78%) laboratory-confirmed cases. This rate does not differ from the rates reported for household or day-care exposure in the same age group. Other k n o w n exposures represent m o r e or less causal contacts with both u n c o n f i r m e d and laboratory-confirmed pertussis. T h e two oldest patients~in this study, with reported exposure to grandchildren, were listed in this group. Both patients had laboratory-confirmed pertussis. T h e rate of laboratory-confirmed disease in this group, however, was not different from the rate among patients without any known exposure.

G. GRANSTROM ET AL.

22

Table II Reported type of exposure in 285 patients with pertussis* and suspected pertussis of other causes Number of laboratory confirmed/number of suspected cases (%) Reported type of exposure

Unimmunised o--i2 months

Unimmunised I-9 years

Immunised older children and adults

Household

5/i2 (42)

Day-care

I/4

School Other known No known

o 4/16 (25) 3/9

I8/20 (90) 89/122 (73) 2/3 9/22 (4 I) II/34 (32)

9/II (82) 0/7 5/6 4/9 3/IO (30)

* Two cases of parapertussis (one day-care and one school exposure) included. Diagnostic value o f findings

T h e prevalence of pertussis in those out-patients seeking medical attention with parental or own suspicion of whooping cough was thus 57 % (I63/285). T h e diagnostic sensitivity and specificity of clinical symptoms and evaluation are shown in Table I. It has to be noted that, for instance, the 37 % sensitivity for whoops cannot be extrapolated for the entire duration of illness. T h e rates are derived from a single first visit made after a week or less of illness by 53 % pertussis patients and by 48 % non-pertussis patients. T h e specificities for whoops and for the two types of physicians' diagnosis were high, i.e. 96, ioo and 99 % respectively, and are less likely to depend on the length of illness. T h e predictive value for the diagnosis of pertussis was thus 64 % for cough with vomiting, 92% for whoops and zoo% for physicians' diagnosis of pertussis. T h e predictive value for diagnosis of non-pertussis was 68 % for fever and 93 % for physicians' diagnosis of non-pertussis. T h e sensitivity and specificity for cough of long duration, i.e. 6 weeks or more, cannot be established exactly, since this information was not obtained for all patients. T h e specificity for this s y m p t o m could in any event not be higher than 84 % (I02/I22). T h e predictive value of epidemiological data by type of reported exposure is directly given in Table II. T h e sensitivity and specificity of epidemiological data may also be calculated from Table II, although this information is less relevant than the predictive value. For instance, the sensitivity of household exposure of 2o % (32/I63) in pertussis patients is significantly different (P o'oi9) from the 9 % (I I/I22) in non-pertussis patients and yields a specificity of 9I % for this type of exposure. T h e highest predictive value for diagnosis of pertussis would thus be obtained in an u n i m m u n i s e d (non-infant) child when the parents report whoops and household exposure and when a physician has diagnosed the disease as pertussis. Bacterial culture

Positive culture for B. pertussis among all patients with positive laboratory findings in relation to duration of illness is shown in Table III. Positive cultures were obtained in 50-60 % patients during the first 2 weeks of disease.

Diagnosis of pertussis

23

Table I I I Sensitivity of culture for Bordetella pertussis* in relation to duration

of illness Duration of illness (weeks)

Number positive cultures/ number laboratory-confirmed cases (%)

~< i 2 3 4

49/86 (57) 22/4I (54) 6/r8 (33) 3/8

5

z/8

>/6 Total

o/2 82/I63 (50)

Two patients with positive B. pertussis culture who refused to give a convalescent blood sample and two patients with positive B. parapertussis culture are included.

Table IV Positive culture for non-Bordetella sp. in nasopharyngeal aspirates

from patients with pertussis* and suspected pertussis of other causes

Bacteria

Branhamella catarrhalis alone Haemophilus influenzae alone Streptococcus pneumoniae alone Combinations of above Total

Pertussis (n = I63) n (%)

Non-pertussis (n = I22) n(%)

39 (24) 22 (I3) 17 (Io) 45 (28) xz3 (75)

27 (22) 14 (I I) I3 (I I) 38 (3I) 92 (75)

* Two cases of parapertussis included.

T h e rate of positive culture decreased to 3o-4 ° % in the following 2 weeks. Culture-positive cases were still found during the 5th week. T h e validity of these findings for other culture media was tested by comparing the B o r d e t - G e n g o u m e d i u m used in the study with charcoal m e d i u m in current use in many laboratories. In a total of Io9 consecutive clinical isolates of B. pertussis from nasopharyngeal swabs, 96 (88 %) isolates grew on both media. In addition, IO were positive on charcoal only and three on B o r d e t - G e n g o u only (a non-significant difference). F o u r B. parapertussis strains were isolated on B o r d e t - G e n g o u medium. T w o of these four isolates did not grow on charcoal agar. Isolates of other potential respiratory pathogens found on culture from the nasopharyngeal aspirates are shown in Table IV. Only some 25 % patients in both groups did not have significant growth. No differences were found between the two groups with regard to distribution of bacteria or combinations of bacteria. T h e total isolation rate for Branhamella catarrhalis was 45 % in the pertussis patients and 53 % in the non-pertussis. T h e corresponding figures for Haemophilus influenzae were 2 9 % and 23% and for Streptococcus pneumoniae 33 % and 39 % respectively.

24

G. GRANSTR()M E T A L .

Discussion

T h e present study reports the diagnostic value of clinical findings and bacterial culture in pertussis with serology as the reference method. Bordetella pertussis was isolated from 50 % patients, confirming its often-quoted sensitivity for culture. 1'2 T h e two media, i.e. B o r d e t - G e n g o u and charcoal (Regan-Lowe) agar, showed a similar performance in the present study. In the original study by Regan and Lowe, 7 claims of superior sensitivity for charcoal m e d i u m were based on the statement that 'the B o r d e t - G e n g o u m e d i u m did not support the growth of the bacterium'. A recent systematic study on laboratory strains showed better results for charcoal agar. 1° In the present study the B o r d e t - G e n g o u plate was streaked first, which could have given this m e d i u m an advantage. T h e main cause of the differences in results is most likely to be the quality of the B o r d e t - G e n g o u agar, a m e d i u m known to be difficult to optimise. Culture for other bacteria in nasopharyngeal aspirates revealed high and equal rates of colonisation in the two groups of patients. T h e results indicate that the presence or absence of these bacteria is of no diagnostic value in confirming or excluding the diagnosis of pertussis. Also, the results do not indicate an important role for these bacteria in pertussis-suspect cough o f n o n Bordetella origin. T h e diagnostic value of the various symptoms and evaluations for clinical diagnosis of pertussis showed that a physician's diagnosis of the disease is the most reliable criterion, followed by the s y m p t o m of whoops. It is of some interest to note that physicians diagnosis attained the same ioo % specificity as culture. T h e results do not therefore sustain the general belief that clinical diagnosis is less reliable. T h e sensitivity of physicians' diagnosis was also the highest among the clinical criteria and was almost identical to the sensitivity of culture. T h e present study was conducted in a mainly u n i m m u n i s e d child population. T h e n u m b e r of immunised children and adults was too small to allow conclusive comparisons between the groups. In this limited group of 2i patients, neither the rate of clinical symptoms nor the physician's diagnosis differed from the rates obtained in the u n i m m u n i s e d children. T h e validity of the results obtained for physicians' diagnosis in a country with such a high incidence as Sweden cannot be established for countries with a low incidence. It might be speculated that physicians are uniikely to become more confident and over-diagnose a disease that is seldom seen and well-known to be difficult to recognise. On the contrary, they would be less likely to venture a diagnosis on clinical symptoms alone. If these assumptions are correct, specificity would be the same but sensitivity would be m u c h lower. T h e epidemiological data on reported exposure showed high predictive values for well-defined exposures in household, day-care and school. It is also of some interest that the predictive value was similar for immunised and u n i m m u n i s e d children over I year of age. Although the data do not represent attack rates, the predictive values indicate that an immunisation schedule with three doses of D T P vaccine given to infants without booster injections, i.e. as earlier practised in Sweden and currently in the U.K., affords no substantial

Diagnosis of pertussis

z5

protection against the disease by the age of 8 years. Of patients without any knowledge of contact, approximately one in three had pertussis, indicating that a lack of known exposure does not exclude the disease. T h e low predictive value of epidemiological data in infants was an unexpected finding. Several reasons and combinations of causes are plausible. One reason could be that parents are familiar with the severe disease that can develop in infants. T h e y are therefore more likely to over-speculate pertussis among the contacts of an infant with respiratory symptoms. This assumption is supported by a large proportion of contacts in the study for whom pertussis could be excluded by laboratory methods. Another reason may be that the serological assays are less likely to identify pertussis correctly in infants. Although this possibility cannot be entirely excluded, it seems less likely, since 12 infants were diagnosed serologically. These included all six culture-positive cases. Neither were there any differences in the rate of timing or treatment with erythromycin, since most infants in both the pertussis and the nonpertussis group received early treatment. T h e cause of infants escaping the acquisition of disease from confirmed contacts could also be protection by maternal antibody. Lower attack rates in infants than in older children were noted in the pre-vaccination era, 11 which resembles the present situation in Sweden, where none of the pre-school children are immunised. Few previous studies are available for comparison with the present results. Although one recent study investigated the diagnostic sensitivity and specificity of clinical symptoms in pertussis, 12 its design differed in that culture was the principal reference method. Neither did it investigate the symptoms and evaluations which had the highest predictive values in the present study. T h e highest predictive value in that study was obtained for cough of at least 14 days duration. 1~ Duration of cough was also extrapolated in the present study to have some predictive value, but both studies found or indicated a specificity below 90 % for this symptom. T h e minimal specificity for a useful diagnostic method depends on the expected prevalence of the disease. As an example, the specificity for serological diagnosis on single samples from the general population is usually set at an arbitrary limit of 95 %. T h e choice is based on the expected low prevalence in the routine clinical situation, which differs from the high prevalence in studies or outbreaks. In conclusion, clinical diagnosis of pertussis by a physician was found to be a highly reliable diagnostic method. T h e only symptom with high specificity was reported whoops. Epidemiological data of well-defined exposure were also of significant value in the diagnosis of the disease. Culture of B. pertussis was found to have an overall sensitivity of 50 %, with comparable results for Bordet-Gengou and charcoal (Regan-Lowe) medium. (This study was supported by contract no. 2oo-84-o752 from the Centers for Disease Control, Atlanta, GA, U.S.A. for the development and evaluation of diagnostic methods for pertussis. We thank Lorita L6fstrand and Kerstin L6vgren for technical assistance. The generous collaboration of our colleagues at the Department of Paediatrics, Danderyd Hospital, made this study possible.)

26

G. GRANSTROM ET AL.

References I. Onorato IM, Wassilak SGF. Laboratory diagnosis of pertussis. State of the art. Pediatr lnfect Dis J I987; 6: I45-I5!. 2. Friedman RL. Pertussis: The disease and new diagnostic methods. Clin Microbiol Rev r988; x: 365-376. 3. Granstr6m M, Granstr6m G, Lindfors, A, Askel6f P. Serologic diagnosis of whooping cough by an enzyme-linked immunosorbent assay using firnbrial hemagglutinin as antigen. J Infect Dis I982; x46: 741-45. 4. Burstyn DG, Baraff LJ, Peppler M S et al. Serologicresponseto filamentous hemagglutinin and lymphocytosis-promoting toxin of Bordetella pertussis. Infect Immunol I983; 4I: I r5o--I I56. 5. Granstr6m G, Wretlind B, Salenstedt CR, Granstr6m M. Evaluation of serologic assays for diagnosis of whooping cough. J Clin Microbiol I988; 26:I818-I823. 6. Grandstr6m G, Askel6f P, Granstr6m M. Specific immunoglobulin A to Bordetella pertussis antigens in mucosal secretion for rapid diagnosis of whooping cough. J Clin Microbiol I988; 26: 869-874. 7. Regan J, Lowe F. Enrichment medium for the isolation of Bordetella. J Clin Microbiol I977; 6: 3o3-309. 8. Gillenius P, J~itmaa E, Askel6f P, Granstr6m M, Tiru M. Standardization of an assay for pertussis toxin and antitoxin in microplate culture of Chinese hamster ovary cells. J Biol Stand r985; x3: 61-66. 9. Granstr6m M, Granstr6m G, Gillenius P, Askel6f P. Neutralizing antibodies to pertussis toxin in whooping cough. J lnfect Dis I985; xSx : 646-649. ro. Hoppe JE, Vogl R. Comparison of three media for culture ofBordetella pertussis. EurJ Clin Microbiol I986; 5: 36r-362. I I. Lapin JH. Whooping cough. Springfield, IL: Charles C. Thomas, I943: 8-24. i2. Patriarca PA, Biellik RJ, Sanden G e t al. Sensitivity and specificity of clinical case definitions for pertussis. Am J Public Health I988; 78:833-836.

Diagnostic value of clinical and bacteriological findings in pertussis.

Clinical and bacteriological findings in the diagnosis of pertussis were evaluated in 300 consecutive patients with parental or the patient's own susp...
685KB Sizes 0 Downloads 0 Views