- Email: [email protected]
Bacterial pathogens of otitis media and sinusitis: Detection in the nasopharynx with selective agar media
Bacterial pathogens of otitis media and sinusitis: Detection in the nasopharynx with selective agar media SHARON DUDLEY, KATHLEEN ASHE, BIRGIT WINTHER, and J. OWEN HENDLEY CHARLOTTESVILLE, VIRGINIA
Carriage rates for the bacterial pathogens associated with otitis media (Streptococcus pneumoniae [SP], Hemophilus influenzae [HI], and Moraxella catarrhalis [MC]) are of interest. Culture on three selective agars was compared with culture on two standard agars to determine the more accurate method for detection of these species in the nasopharynx of healthy children. Weekly samples were obtained in winter from 18 healthy children (ages 1 through 9 years) as part of a longitudinal study. A 0.1-mL sample of 116 nasopharyngeal aspirate/washes was inoculated onto each of five agars. Two were standard (sheep blood and chocolate), and three were selective (blood with gentamicin for SP; chocolate with vancomycin, bacitracin, and clindamycin for HI; blood with amphotericin B, vancomycin, trimethoprim, and acetazolamide for MC). One technician read the standard plates and another the selective; both were blinded to the results of the other. SP was found in 44% of samples with selective agar versus 25% with standard agar; HI was found in 31% with selective versus 9% with standard; MC was found in 56% with selective versus 37% with standard. Overall, 80% of samples had one or more pathogens detected with selective agars as compared with 58% with standard agars (P = .0004). Selective agars were more accurate than standard agars for detecting otitis pathogens in the nasopharynx, where they are a common part of normal flora in healthy children. (J Lab Clin Med 2001;138:338-42)
Abbreviations: CFU = colony-forming unit; HI = Hemophilus influenzae; MC = Moraxella catarrhalis; SP = Streptococcus pneumoniae
T
he bacterial strains SP, HI, and MC cause bacterial otitis media in children and bacterial sinusitis in adults and children. The ability to accurately detect the presence of these bacterial species in the
From the Department of Clinical Microbiology; Department of Pediatrics; and Department of Otolaryngology, Head and Neck Surgery, University of Virginia Health System. Supported by The Pendleton Pediatric Infectious Disease Laboratory at the University of Virginia Health System. Submitted for publication April 9, 2001; revision submitted August 3, 2001; accepted August 8, 2001. Reprint requests: J. Owen Hendley, MD, University of Virginia Health System, Department of Pediatrics, Box 800386, Division of Pediatric Infectious Disease, Charlottesville, VA 22908. Copyright © 2001 by Mosby, Inc. 0022-2143/2001 $35.00 + 0 5/1/119311 doi:10.1067/mlc.2001.119311 338
nasopharynx is of interest,1-4 particularly when studies are evaluating potential prevention strategies. The two components that are important in accurate detection of the organisms in the nasopharynx are the method by which the sample of secretions is obtained and the means for detecting the presence of the organisms in the sample. Secretions have been sampled by nasal swab, nasal wash, nasal aspirate, and oral swab.5-8 Once the sample has been obtained, the major problem in detection is overgrowth of the bacterial species of interest by other bacteria present in the nasopharynx or acquired in the process of obtaining the sample. Antibiotics and other inhibitors have been incorporated into culture media to prevent overgrowth of obscuring bacteria. Gentamicin has been included in blood agar for detection of SP.9 Bacitracin has been included in an improved chocolate agar for detection of HI.10 Chapin and Doern11 reported that the addition of vancomycin
J Lab Clin Med Volume 138, Number 5
and clindamycin to bacitracin was helpful. The addition of acetazolamide to three antibiotics was shown by Belgian workers12 to enhance the detection of MC by inhibiting the growth of Neisseria species. In an ongoing longitudinal study, we wish to detect all three of these respiratory bacterial pathogens in nasopharyngeal samples. We obtain samples by an aspirate/wash method that is tolerable for repeated weekly sampling in healthy children in the United States. In the current study, we examined the sensitivity of methods for detection of the three pathogens in the samples. We compared the use of two non-selective agar media (sheep blood agar and enriched chocolate agar) to results with three selective agars, each designed to detect one of the bacterial pathogens. METHODS Subjects and sampling. Eighteen healthy children living in the Charlottesville community participating in an ongoing study of the effect of viral respiratory infection on the middle ear have a sample of nasopharyngeal secretions obtained once a week. The study was approved by the Human Investigation Committee at the University of Virginia and parents signed informed consent. Fourteen children are ages 1 through 5 years and 4 children are ages 6 to 9 years. Samples obtained from mid January through early March of 1999 were used in this study. During this 8-week period, an upper respiratory illness occurred in all 18 children; 3 children received an antibiotic, and 2 of these 3 were diagnosed with otitis media. One child could not be sampled successfully and left the study. The number of samples from each child ranged from 2 to 8. The mean number of samples per child was 6. The nasopharyngeal aspirate/wash was collected with a Yankauer suction device (Conmed, Utica, NY) placed in the nasal vestibule and connected to a Lukens specimen trap (Sherwood Medical, St Louis, MO) with a portable suction pump (Vacu-aide, –90 mm Hg; DeVilbiss, Sunrise Medical Co, Somerset, PA). Suction was applied for 10 to 15 seconds in each nostril. If visible secretion was not aspirated, 1 to 2 mL of saline solution (sodium chloride inhalation solution USP; Dey Laboratories, Napa, CA) was dropped into one nostril followed by aspiration from the contralateral nostril. Secretion in the tubing was then washed into the Lukens trap by aspiration of the remainder of the 5-mL aliquot of saline solution. The material in the trap, designated nasopharyngeal aspirate/wash, was mixed thoroughly and used to inoculate agar plates for detection of bacteria. Standard and selective agar systems. The two agar plates that comprised the standard (non-selective) system were 5% sheep blood in trypticase soy agar (TSA II, 5% sheep blood) and enriched chocolate agar (Choc II), both of which were obtained as prepared media (BBL; Becton-Dickinson, Cockeysville, MD). The three agar plates in the selective system were made in our laboratory. An 18-mL sample of each agar was contained in 100 × 15 mm Petri dishes. For SP, 5% sheep blood in tryptone agar base (BBL) with 5 µg gentamicin per milliliter was used.9 For HI, 5% chocolated sheep blood in GC agar base
Dudley et al
339
(BBL) with 1% yeast autolysate (Oxoid, Basingstoke, England) was used.10 Vancomycin (5 µgm/mL), bacitracin (300 µg/mL), and clindamycin (1 µg/mL) were included in the agar, as described by Chapin and Doern.11 For MC, brucella agar base (BBL) with 5% sheep blood was used, with the inclusion of amphotericin B (2 µgm/mL), vancomycin (10 µg/mL), trimethoprim (5 µgm/mL), and acetazolamide (10 µgm/mL) as described by Vaneechouette et al.12 Inoculated agar plates for MC were incubated in air, and the other four plates were incubated in a candle extinction jar at 35°C for 48 hours. Inoculation of agar and identification of organisms. A 0.1-mL aliquot of nasopharyngeal aspirate/wash was inoculated onto each of the five agars. The inoculum on selective agars was dispersed evenly over the plate by using a loop; on standard agars it was spread with a loop for the usual threeway streak, because overgrowth by other flora made the plates almost unreadable if the inoculum was evenly dispersed. Quantitation of growth of the pathogen in CFU/0.1 mL on selective agars was provided by counting the number of colonies on the entire plate on which 0.1 mL of aspirate/wash had been spread. On non-selective agars, a semiquantitative estimate of the number of CFU was determined by grading the growth of each pathogen as 1+ (<10 CFU), 2+ (10 to 50 CFU), 3+ (50 to 100 CFU), or 4+ (>100 CFU) based on the detection of colonies of the organism in first, second, third, or fourth streak, respectively. An ASCP-certified technologist with 30 years of experience in clinical microbiology read the standard system plates, and a research technician with 5 years of experience in identifying respiratory bacteria read the selective plates. The nonselective plates were read in one room and the selective plates in another room. There was no shared reading of plates or cross-checking; results were not compared until the time of analysis after the study period had ended. Colonies of the three bacterial species were picked and identified by established methods. SP was identified by sensitivity to optochin (P disc, BBL); HI was identified on trypticase soy agar with paper disks impregnated with nicotinamide adenine dinucleotide, hemin, or both (factors X, V, and XV; Difco, Detroit, MI). MC was identified by gram stain, oxidase positivity, and positive indoxyl acetate strip test as described by Speeleveld et al.13 Statistical analysis. Differences in detection frequency were examined with the two-tailed Fisher exact test. RESULTS
At least one of the three bacterial species was detected in 98 (84%) of the 116 nasopharyngeal aspirate/wash samples obtained in the winter during weekly sampling of the 18 children (Tables I and II). One species was detected in 54 (47%) samples, two species were found in 25 (22%) samples, and all three species were detected in 19 (16%) of the 116 samples. Ninety-five percent of the positive samples were detected with the selective agar system (Table I), as compared with 68% of positives detected with the standard system (P < .001). SP and MC were detected on standard agar in 57% and
J Lab Clin Med November 2001
340 Dudley et al
Table I. Bacterial pathogens in 116 nasopharyngeal aspirate/wash cultures of healthy children 1 to 9 years of age: Proportion of positive samples* detected on selective versus standard agars Bacterial species
Number of positive samples
S pneumoniae H influenzae M catarrhalis Any pathogen
Selective agar
51 39 71 98
50 37 66 93
Standard agar
(98%) (95%) (93%) (95%)
29 10 43 67
(57%) (26%) (61%) (68%)
*Cultures in which one of the three pathogens was detected with either agar system inoculated with 0.1 mL of nasopharyngeal aspirate/wash.
Table II. Bacterial pathogens in 116 nasopharyngeal aspirate/wash cultures of healthy children 1 to 9 years of age: Carriage rates determined by selective versus standard agars Selective agar Number positive (rate)
Bacterial species
S pneumoniae H influenzae M catarrhalis Any pathogen
50 37 66 93
Standard Agar Number positive (rate)
(43%) (32%) (57%) (80%)
29 10 43 67
P
(25%) (9%) (37%) (58%)
.005 .0002 .04 .0004
Table III. Bacterial pathogens in 116 nasopharyngeal aspirate/wash cultures: Effect of titer of pathogen in sample on detection rate on standard agar S pneumoniae
H influenzae
M catarrhalis
Selective number
Standard number (rate]
Selective number
Standard number (rate]
Selective number
Standard number (rate]
≥102 101 to 102 100 to 101 <100†
25 9 16 1
23 (93%) 4 (44%) 1 (16%)
8 7 22 2
6 (75%) 1 (14%) 1( 5%)
33 20 13 5
24 (73%) 10 (50%) 4 (31%)
Total positive
51
Titer*
39
71
*Titer on selective agar expressed in CFU/0.1 mL. †Not detected on selective agar.
61%, respectively, of the samples containing those species; HI was detected with standard agar in only 26% of samples in which it was present. The organism was detected with standard agar in 73% to 93% of samples containing high titers (102 CFU/0.1 mL) of pathogen as ascertained on selective agar (Table III). The detection rate with standard agar dropped precipitously in samples with titers of pathogen lower than 102 CFU/0.1 mL. HI was particularly notable, with a 5% detection rate on standard chocolate agar in 22 samples containing 1 to 10 CFU/0.1 mL on selective agar. Inhibition of growth of the pathogen on the selective agars did not appear to be a problem: only 1/51 (2%) of SP isolates, 2/39 (5%) of HI isolates, and 5/71 (7%) of isolates of MC were detected on standard agar and not on selective agar.
The carriage rates of each of three pathogens in nasopharyngeal aspirate/wash samples as estimated with the selective system were obviously greater than rates estimated with the standard agar system (Table II). DISCUSSION
The picture of nasopharyngeal carriage of these pathogenic bacteria is different based purely on the isolation method used to detect their presence in the nasopharyngeal secretion samples. The three selective agars provided a much more accurate assessment of the presence of the pathogens than did the nonselective agars. Very few isolates were found only with standard agars and not with selective. The enhanced sensitivity of detection with the selective system was
J Lab Clin Med Volume 138, Number 5
obviously due to inhibition of overgrowth of other flora (Fig 1), allowing detection of the presence of the pathogen even in low titer. All but 8 of the 161 isolates of one of the three pathogens were detected on the selective agars. Of the 8 isolates missed with the selective agar, 1 isolate of MC had 50 to 100 colonies (3+) on standard agar, 3 isolates (1 of each species) had 10 to 50 colonies (2+) on standard agar, and 4 isolates (3 of MC and 1 of HI) had 1 to 10 colonies. It is possible that the antibiotics/inhibitors in the selective agars were preventing detection of the species of interest, but a more likely explanation for the missed isolates on selective agar is inhomogenity of the sample. The selective agars used in this study were developed by others. The addition of gentamicin to sheep blood agar for SP was described by Black and Buskirk14 and evaluated by Converse and Dillon.9 Chapin and Doern11 noted frequent breakthroughs of other flora when bacitracin alone was used in selective agar for HI, which led to inclusion of the three antibiotics. In addition, we used an optimized agar formulation described by Rennie et al in this study.10 For MC, a Belgian group12 described the addition of acetazolamide, a carbonic anhydrase inhibitor, to agar containing three antibiotics that select for MC and Neisseria species. Neisseria species commonly produce carbonic anhydrase, which is essential for growth of the organism; MC does not produce it.12 Inhibition of the enzyme by acetazolamide can be overcome in an atmosphere rich in CO2. Vaneechoutte et al12 demonstrated that incubation of agar plates in air resulted in inhibition of Neisseria species, which allowed detection of MC. These selective agars have been shown to enhance detection of the pathogens in respiratory tract samples including nasopharyngeal secretions, throat cultures, and sputum samples. Converse and Dillon9 demonstrated that the addition of gentamicin to blood agar increased the detection rate of SP from 33% to 38% in 86 nasopharyngeal swab samples and from 23% to 37% in 84 throat swab samples from infants. Gentamicin agar has been utilized for pneumococcal detection in recent carriage studies.5,6 For detection of HI in throat swab samples from 852 children in daycare, Chapin and Doern 11 compared non-selective chocolate agar to the selective agar with vancomycin, bacitracin, and clindamycin. They demonstrated that HI was detected in only 16% of throat samples with chocolate agar, versus 60% with selective agar. MC was detected with routine blood agar in about 6 out of 10 samples of respiratory tract secretions in which its presence was demonstrated with the selective agar by Vaneechouette and coworkers. 15 They noted that the increased isolation rate with selective agar was pri-
Dudley et al
341
Fig 1. Streptococcus pneumoniae in 0.1 mL of a nasopharyngeal aspirate/wash inoculated onto standard blood agar (left panel) or selective agar (right panel). Alpha hemolytic colonies (≥102 CFU/0.1 mL) of the organism are obvious on selective agar.
marily due to the detection of <20 colonies in the inoculum, which is in accord with our findings. The three selective agars are not commercially available, but the cost including ingredients and technical time for making the plates is modest. In this study, plates were made in batches for about $0.60 apiece, or $1.80 for all three plates. The two non-selective agar plates were purchased at a discounted price of $0.25 each (or $0.62 retail). The total price per sample tested was $1.80 for selective media and $0.50 (discounted) or $1.24 (retail) for non-selective media. The ease of reading the selective plates makes them attractive in spite of the price difference. The sampling method we used in this study is tolerable for weekly use in small children, and different methods for obtaining samples of nasopharyngeal secretions were not examined. When coupled with determination of the dilution of “nasal lining fluid” by measurement of the urea concentration,16 the selective agars provide quantitative assessment of pathogens in nasopharyngeal secretions. Regardless of how the sample is obtained, the inoculation of 0.1 mL of sample onto selective agars is a reliable means of determining the presence and titer of the organisms. It will be of interest to see whether titers are different in children at the time of bacterial otitis media2 or in adults with colds.1 REFERENCES
1. Kaiser L, Lew D, Hirschel B, Auckenthaler R, Morabia A, Heald A, et al. Effects of antibiotic treatment in the subset of common-cold patients who have bacteria in nasopharyngeal secretions. Lancet 1996;347:1507-10.
342 Dudley et al
2. Faden H, Stanievich J, Brodsky L, Bernstein J, Ogra PL. Changes in nasopharyngeal flora during otitis media in childhood. Pediatr Infect Dis J 1990;9:623-6. 3. Prellner K, Christensen P, Hovelius B, Rosen C. Nasopharyngeal carriage of bacteria in otitis-prone and non-otitis prone children in day-care centres. Acta Otolaryngol 1984; 98:343-50. 4. Faden H, Waz MJ, Bernstein JM, Brodskey L, Stanievich J, Ogra PL. Naospharyngeal flora in the first three years of life in normal and otitis-prone children. Ann Otol Rhinol Laryngol 1991;100:612-5. 5. Capeding MR, Nohynek H, Sombrero LT, Pascual LG, Sunico ES, Esparar GA, et al. Evaluation of sampling sites for detection of upper respiratory tract carriage of Streptococcus pneumoniae and Haemophilus influenzae among healthy Filipino infants. J Clin Microbiol 1995;33: 3077-9. 6. Rapola S, Salo E, Kiiski P, Leinonen M, Takala AK. Comparison of four different sampling methods for detecting pharyngeal carriage of Streptococcus pneumoniae and Haemophilus influenzae in children. J Clin Microbiol 1997; 35:1077-9. 7. Lankinen KS, Leiononen M, Tupasi TE, Haikala R, Ruutu P. Pneumococci in nasopharyngeal samples from Filipino children with acute respiratory infections. J Clin Microbiol 1994;32:2948-52. 8. Homøe P, Prag J, Farholt S, Henrichsen J, Hornsleth A, Kilian M, et al. High rate of nasopharyngeal carriage of potential pathogens among children in Greenland: results of a clinical survey of middle-ear disease. Clin Infect Dis 1996; 23:1081-90.
J Lab Clin Med November 2001
9. Converse GM III, Dillon HC Jr. Epidemiological studies of Streptococcus pneumoniae in infants: methods of isolating pneumococci. J Clin Microbiol 1977;5:293-6. 10. Rennie R, Gordon T, Yaschuk Y, Tomlin P, Kibsey P, Albritton W. Laboratory and clinical evaluations of media for the primary isolation of Haemophilus species. J Clin Microbiol 1992;30:1917-21. 11. Chapin KC, Doern GV. Selective media for recovery of Haemophilus influenzae from specimens contaminated with upper respiratory tract microbial flora. J Clin Microbiol 1983;17:1163-5. 12. Vaneechoutte M, Verschraegen G, Claeys G, van den Abeele AM. Selective medium for Branhamella catarrhalis with acetazolamide as a specific inhibitor of Neisseria spp. J Clin Microbiol 1988;26:2544-8. 13. Speeleveld E, Fossépré J-M, Gordts B, Van Landuyt HW. Comparison of three rapid methods, tributyrine, 4-methylumbelliferyl butyrate, and indoxyl acetate for rapid identification of Moraxella catarrhalis. J Clin Microbiol 1994;32: 1362-3. 14. Black WA, Buskirk FV. Gentamicin blood agar as a generalpurpose selective medium. Appl Microbiol 1973;25:905-7. 15. Vaneechoutte M, Verschraegen G, Claeys G, Weise B, van den Abeele AM. Respiratory tract carrier rates of Moraxella (Branhamella) catarrhalis in adults and children and interpretation of the isolation of M. catarrhalis from sputum. J Clin Microbiol 1990;28:2674-80. 16. Kaulbach HC, White MV, Igarashi Y, Hahn BK, Kaliner MA. Estimation of nasal epithelial lining fluid using urea as a marker. J Allergy Clin Immunol 1993;92:457-65.
Carriage rates for the bacterial pathogens associated with otitis media (Streptococcus pneumoniae [SP], Hemophilus influenzae [HI], and Moraxella catarrhalis [MC]) are of interest. Culture on three selective agars was compared with culture on two standard agars to determine the more accurate method for detection of these species in the nasopharynx of healthy children. Weekly samples were obtained in winter from 18 healthy children (ages 1 through 9 years) as part of a longitudinal study. A 0.1-mL sample of 116 nasopharyngeal aspirate/washes was inoculated onto each of five agars. Two were standard (sheep blood and chocolate), and three were selective (blood with gentamicin for SP; chocolate with vancomycin, bacitracin, and clindamycin for HI; blood with amphotericin B, vancomycin, trimethoprim, and acetazolamide for MC). One technician read the standard plates and another the selective; both were blinded to the results of the other. SP was found in 44% of samples with selective agar versus 25% with standard agar; HI was found in 31% with selective versus 9% with standard; MC was found in 56% with selective versus 37% with standard. Overall, 80% of samples had one or more pathogens detected with selective agars as compared with 58% with standard agars (P = .0004). Selective agars were more accurate than standard agars for detecting otitis pathogens in the nasopharynx, where they are a common part of normal flora in healthy children. (J Lab Clin Med 2001;138:338-42)
Abbreviations: CFU = colony-forming unit; HI = Hemophilus influenzae; MC = Moraxella catarrhalis; SP = Streptococcus pneumoniae
T
he bacterial strains SP, HI, and MC cause bacterial otitis media in children and bacterial sinusitis in adults and children. The ability to accurately detect the presence of these bacterial species in the
From the Department of Clinical Microbiology; Department of Pediatrics; and Department of Otolaryngology, Head and Neck Surgery, University of Virginia Health System. Supported by The Pendleton Pediatric Infectious Disease Laboratory at the University of Virginia Health System. Submitted for publication April 9, 2001; revision submitted August 3, 2001; accepted August 8, 2001. Reprint requests: J. Owen Hendley, MD, University of Virginia Health System, Department of Pediatrics, Box 800386, Division of Pediatric Infectious Disease, Charlottesville, VA 22908. Copyright © 2001 by Mosby, Inc. 0022-2143/2001 $35.00 + 0 5/1/119311 doi:10.1067/mlc.2001.119311 338
nasopharynx is of interest,1-4 particularly when studies are evaluating potential prevention strategies. The two components that are important in accurate detection of the organisms in the nasopharynx are the method by which the sample of secretions is obtained and the means for detecting the presence of the organisms in the sample. Secretions have been sampled by nasal swab, nasal wash, nasal aspirate, and oral swab.5-8 Once the sample has been obtained, the major problem in detection is overgrowth of the bacterial species of interest by other bacteria present in the nasopharynx or acquired in the process of obtaining the sample. Antibiotics and other inhibitors have been incorporated into culture media to prevent overgrowth of obscuring bacteria. Gentamicin has been included in blood agar for detection of SP.9 Bacitracin has been included in an improved chocolate agar for detection of HI.10 Chapin and Doern11 reported that the addition of vancomycin
J Lab Clin Med Volume 138, Number 5
and clindamycin to bacitracin was helpful. The addition of acetazolamide to three antibiotics was shown by Belgian workers12 to enhance the detection of MC by inhibiting the growth of Neisseria species. In an ongoing longitudinal study, we wish to detect all three of these respiratory bacterial pathogens in nasopharyngeal samples. We obtain samples by an aspirate/wash method that is tolerable for repeated weekly sampling in healthy children in the United States. In the current study, we examined the sensitivity of methods for detection of the three pathogens in the samples. We compared the use of two non-selective agar media (sheep blood agar and enriched chocolate agar) to results with three selective agars, each designed to detect one of the bacterial pathogens. METHODS Subjects and sampling. Eighteen healthy children living in the Charlottesville community participating in an ongoing study of the effect of viral respiratory infection on the middle ear have a sample of nasopharyngeal secretions obtained once a week. The study was approved by the Human Investigation Committee at the University of Virginia and parents signed informed consent. Fourteen children are ages 1 through 5 years and 4 children are ages 6 to 9 years. Samples obtained from mid January through early March of 1999 were used in this study. During this 8-week period, an upper respiratory illness occurred in all 18 children; 3 children received an antibiotic, and 2 of these 3 were diagnosed with otitis media. One child could not be sampled successfully and left the study. The number of samples from each child ranged from 2 to 8. The mean number of samples per child was 6. The nasopharyngeal aspirate/wash was collected with a Yankauer suction device (Conmed, Utica, NY) placed in the nasal vestibule and connected to a Lukens specimen trap (Sherwood Medical, St Louis, MO) with a portable suction pump (Vacu-aide, –90 mm Hg; DeVilbiss, Sunrise Medical Co, Somerset, PA). Suction was applied for 10 to 15 seconds in each nostril. If visible secretion was not aspirated, 1 to 2 mL of saline solution (sodium chloride inhalation solution USP; Dey Laboratories, Napa, CA) was dropped into one nostril followed by aspiration from the contralateral nostril. Secretion in the tubing was then washed into the Lukens trap by aspiration of the remainder of the 5-mL aliquot of saline solution. The material in the trap, designated nasopharyngeal aspirate/wash, was mixed thoroughly and used to inoculate agar plates for detection of bacteria. Standard and selective agar systems. The two agar plates that comprised the standard (non-selective) system were 5% sheep blood in trypticase soy agar (TSA II, 5% sheep blood) and enriched chocolate agar (Choc II), both of which were obtained as prepared media (BBL; Becton-Dickinson, Cockeysville, MD). The three agar plates in the selective system were made in our laboratory. An 18-mL sample of each agar was contained in 100 × 15 mm Petri dishes. For SP, 5% sheep blood in tryptone agar base (BBL) with 5 µg gentamicin per milliliter was used.9 For HI, 5% chocolated sheep blood in GC agar base
Dudley et al
339
(BBL) with 1% yeast autolysate (Oxoid, Basingstoke, England) was used.10 Vancomycin (5 µgm/mL), bacitracin (300 µg/mL), and clindamycin (1 µg/mL) were included in the agar, as described by Chapin and Doern.11 For MC, brucella agar base (BBL) with 5% sheep blood was used, with the inclusion of amphotericin B (2 µgm/mL), vancomycin (10 µg/mL), trimethoprim (5 µgm/mL), and acetazolamide (10 µgm/mL) as described by Vaneechouette et al.12 Inoculated agar plates for MC were incubated in air, and the other four plates were incubated in a candle extinction jar at 35°C for 48 hours. Inoculation of agar and identification of organisms. A 0.1-mL aliquot of nasopharyngeal aspirate/wash was inoculated onto each of the five agars. The inoculum on selective agars was dispersed evenly over the plate by using a loop; on standard agars it was spread with a loop for the usual threeway streak, because overgrowth by other flora made the plates almost unreadable if the inoculum was evenly dispersed. Quantitation of growth of the pathogen in CFU/0.1 mL on selective agars was provided by counting the number of colonies on the entire plate on which 0.1 mL of aspirate/wash had been spread. On non-selective agars, a semiquantitative estimate of the number of CFU was determined by grading the growth of each pathogen as 1+ (<10 CFU), 2+ (10 to 50 CFU), 3+ (50 to 100 CFU), or 4+ (>100 CFU) based on the detection of colonies of the organism in first, second, third, or fourth streak, respectively. An ASCP-certified technologist with 30 years of experience in clinical microbiology read the standard system plates, and a research technician with 5 years of experience in identifying respiratory bacteria read the selective plates. The nonselective plates were read in one room and the selective plates in another room. There was no shared reading of plates or cross-checking; results were not compared until the time of analysis after the study period had ended. Colonies of the three bacterial species were picked and identified by established methods. SP was identified by sensitivity to optochin (P disc, BBL); HI was identified on trypticase soy agar with paper disks impregnated with nicotinamide adenine dinucleotide, hemin, or both (factors X, V, and XV; Difco, Detroit, MI). MC was identified by gram stain, oxidase positivity, and positive indoxyl acetate strip test as described by Speeleveld et al.13 Statistical analysis. Differences in detection frequency were examined with the two-tailed Fisher exact test. RESULTS
At least one of the three bacterial species was detected in 98 (84%) of the 116 nasopharyngeal aspirate/wash samples obtained in the winter during weekly sampling of the 18 children (Tables I and II). One species was detected in 54 (47%) samples, two species were found in 25 (22%) samples, and all three species were detected in 19 (16%) of the 116 samples. Ninety-five percent of the positive samples were detected with the selective agar system (Table I), as compared with 68% of positives detected with the standard system (P < .001). SP and MC were detected on standard agar in 57% and
J Lab Clin Med November 2001
340 Dudley et al
Table I. Bacterial pathogens in 116 nasopharyngeal aspirate/wash cultures of healthy children 1 to 9 years of age: Proportion of positive samples* detected on selective versus standard agars Bacterial species
Number of positive samples
S pneumoniae H influenzae M catarrhalis Any pathogen
Selective agar
51 39 71 98
50 37 66 93
Standard agar
(98%) (95%) (93%) (95%)
29 10 43 67
(57%) (26%) (61%) (68%)
*Cultures in which one of the three pathogens was detected with either agar system inoculated with 0.1 mL of nasopharyngeal aspirate/wash.
Table II. Bacterial pathogens in 116 nasopharyngeal aspirate/wash cultures of healthy children 1 to 9 years of age: Carriage rates determined by selective versus standard agars Selective agar Number positive (rate)
Bacterial species
S pneumoniae H influenzae M catarrhalis Any pathogen
50 37 66 93
Standard Agar Number positive (rate)
(43%) (32%) (57%) (80%)
29 10 43 67
P
(25%) (9%) (37%) (58%)
.005 .0002 .04 .0004
Table III. Bacterial pathogens in 116 nasopharyngeal aspirate/wash cultures: Effect of titer of pathogen in sample on detection rate on standard agar S pneumoniae
H influenzae
M catarrhalis
Selective number
Standard number (rate]
Selective number
Standard number (rate]
Selective number
Standard number (rate]
≥102 101 to 102 100 to 101 <100†
25 9 16 1
23 (93%) 4 (44%) 1 (16%)
8 7 22 2
6 (75%) 1 (14%) 1( 5%)
33 20 13 5
24 (73%) 10 (50%) 4 (31%)
Total positive
51
Titer*
39
71
*Titer on selective agar expressed in CFU/0.1 mL. †Not detected on selective agar.
61%, respectively, of the samples containing those species; HI was detected with standard agar in only 26% of samples in which it was present. The organism was detected with standard agar in 73% to 93% of samples containing high titers (102 CFU/0.1 mL) of pathogen as ascertained on selective agar (Table III). The detection rate with standard agar dropped precipitously in samples with titers of pathogen lower than 102 CFU/0.1 mL. HI was particularly notable, with a 5% detection rate on standard chocolate agar in 22 samples containing 1 to 10 CFU/0.1 mL on selective agar. Inhibition of growth of the pathogen on the selective agars did not appear to be a problem: only 1/51 (2%) of SP isolates, 2/39 (5%) of HI isolates, and 5/71 (7%) of isolates of MC were detected on standard agar and not on selective agar.
The carriage rates of each of three pathogens in nasopharyngeal aspirate/wash samples as estimated with the selective system were obviously greater than rates estimated with the standard agar system (Table II). DISCUSSION
The picture of nasopharyngeal carriage of these pathogenic bacteria is different based purely on the isolation method used to detect their presence in the nasopharyngeal secretion samples. The three selective agars provided a much more accurate assessment of the presence of the pathogens than did the nonselective agars. Very few isolates were found only with standard agars and not with selective. The enhanced sensitivity of detection with the selective system was
J Lab Clin Med Volume 138, Number 5
obviously due to inhibition of overgrowth of other flora (Fig 1), allowing detection of the presence of the pathogen even in low titer. All but 8 of the 161 isolates of one of the three pathogens were detected on the selective agars. Of the 8 isolates missed with the selective agar, 1 isolate of MC had 50 to 100 colonies (3+) on standard agar, 3 isolates (1 of each species) had 10 to 50 colonies (2+) on standard agar, and 4 isolates (3 of MC and 1 of HI) had 1 to 10 colonies. It is possible that the antibiotics/inhibitors in the selective agars were preventing detection of the species of interest, but a more likely explanation for the missed isolates on selective agar is inhomogenity of the sample. The selective agars used in this study were developed by others. The addition of gentamicin to sheep blood agar for SP was described by Black and Buskirk14 and evaluated by Converse and Dillon.9 Chapin and Doern11 noted frequent breakthroughs of other flora when bacitracin alone was used in selective agar for HI, which led to inclusion of the three antibiotics. In addition, we used an optimized agar formulation described by Rennie et al in this study.10 For MC, a Belgian group12 described the addition of acetazolamide, a carbonic anhydrase inhibitor, to agar containing three antibiotics that select for MC and Neisseria species. Neisseria species commonly produce carbonic anhydrase, which is essential for growth of the organism; MC does not produce it.12 Inhibition of the enzyme by acetazolamide can be overcome in an atmosphere rich in CO2. Vaneechoutte et al12 demonstrated that incubation of agar plates in air resulted in inhibition of Neisseria species, which allowed detection of MC. These selective agars have been shown to enhance detection of the pathogens in respiratory tract samples including nasopharyngeal secretions, throat cultures, and sputum samples. Converse and Dillon9 demonstrated that the addition of gentamicin to blood agar increased the detection rate of SP from 33% to 38% in 86 nasopharyngeal swab samples and from 23% to 37% in 84 throat swab samples from infants. Gentamicin agar has been utilized for pneumococcal detection in recent carriage studies.5,6 For detection of HI in throat swab samples from 852 children in daycare, Chapin and Doern 11 compared non-selective chocolate agar to the selective agar with vancomycin, bacitracin, and clindamycin. They demonstrated that HI was detected in only 16% of throat samples with chocolate agar, versus 60% with selective agar. MC was detected with routine blood agar in about 6 out of 10 samples of respiratory tract secretions in which its presence was demonstrated with the selective agar by Vaneechouette and coworkers. 15 They noted that the increased isolation rate with selective agar was pri-
Dudley et al
341
Fig 1. Streptococcus pneumoniae in 0.1 mL of a nasopharyngeal aspirate/wash inoculated onto standard blood agar (left panel) or selective agar (right panel). Alpha hemolytic colonies (≥102 CFU/0.1 mL) of the organism are obvious on selective agar.
marily due to the detection of <20 colonies in the inoculum, which is in accord with our findings. The three selective agars are not commercially available, but the cost including ingredients and technical time for making the plates is modest. In this study, plates were made in batches for about $0.60 apiece, or $1.80 for all three plates. The two non-selective agar plates were purchased at a discounted price of $0.25 each (or $0.62 retail). The total price per sample tested was $1.80 for selective media and $0.50 (discounted) or $1.24 (retail) for non-selective media. The ease of reading the selective plates makes them attractive in spite of the price difference. The sampling method we used in this study is tolerable for weekly use in small children, and different methods for obtaining samples of nasopharyngeal secretions were not examined. When coupled with determination of the dilution of “nasal lining fluid” by measurement of the urea concentration,16 the selective agars provide quantitative assessment of pathogens in nasopharyngeal secretions. Regardless of how the sample is obtained, the inoculation of 0.1 mL of sample onto selective agars is a reliable means of determining the presence and titer of the organisms. It will be of interest to see whether titers are different in children at the time of bacterial otitis media2 or in adults with colds.1 REFERENCES
1. Kaiser L, Lew D, Hirschel B, Auckenthaler R, Morabia A, Heald A, et al. Effects of antibiotic treatment in the subset of common-cold patients who have bacteria in nasopharyngeal secretions. Lancet 1996;347:1507-10.
342 Dudley et al
2. Faden H, Stanievich J, Brodsky L, Bernstein J, Ogra PL. Changes in nasopharyngeal flora during otitis media in childhood. Pediatr Infect Dis J 1990;9:623-6. 3. Prellner K, Christensen P, Hovelius B, Rosen C. Nasopharyngeal carriage of bacteria in otitis-prone and non-otitis prone children in day-care centres. Acta Otolaryngol 1984; 98:343-50. 4. Faden H, Waz MJ, Bernstein JM, Brodskey L, Stanievich J, Ogra PL. Naospharyngeal flora in the first three years of life in normal and otitis-prone children. Ann Otol Rhinol Laryngol 1991;100:612-5. 5. Capeding MR, Nohynek H, Sombrero LT, Pascual LG, Sunico ES, Esparar GA, et al. Evaluation of sampling sites for detection of upper respiratory tract carriage of Streptococcus pneumoniae and Haemophilus influenzae among healthy Filipino infants. J Clin Microbiol 1995;33: 3077-9. 6. Rapola S, Salo E, Kiiski P, Leinonen M, Takala AK. Comparison of four different sampling methods for detecting pharyngeal carriage of Streptococcus pneumoniae and Haemophilus influenzae in children. J Clin Microbiol 1997; 35:1077-9. 7. Lankinen KS, Leiononen M, Tupasi TE, Haikala R, Ruutu P. Pneumococci in nasopharyngeal samples from Filipino children with acute respiratory infections. J Clin Microbiol 1994;32:2948-52. 8. Homøe P, Prag J, Farholt S, Henrichsen J, Hornsleth A, Kilian M, et al. High rate of nasopharyngeal carriage of potential pathogens among children in Greenland: results of a clinical survey of middle-ear disease. Clin Infect Dis 1996; 23:1081-90.
J Lab Clin Med November 2001
9. Converse GM III, Dillon HC Jr. Epidemiological studies of Streptococcus pneumoniae in infants: methods of isolating pneumococci. J Clin Microbiol 1977;5:293-6. 10. Rennie R, Gordon T, Yaschuk Y, Tomlin P, Kibsey P, Albritton W. Laboratory and clinical evaluations of media for the primary isolation of Haemophilus species. J Clin Microbiol 1992;30:1917-21. 11. Chapin KC, Doern GV. Selective media for recovery of Haemophilus influenzae from specimens contaminated with upper respiratory tract microbial flora. J Clin Microbiol 1983;17:1163-5. 12. Vaneechoutte M, Verschraegen G, Claeys G, van den Abeele AM. Selective medium for Branhamella catarrhalis with acetazolamide as a specific inhibitor of Neisseria spp. J Clin Microbiol 1988;26:2544-8. 13. Speeleveld E, Fossépré J-M, Gordts B, Van Landuyt HW. Comparison of three rapid methods, tributyrine, 4-methylumbelliferyl butyrate, and indoxyl acetate for rapid identification of Moraxella catarrhalis. J Clin Microbiol 1994;32: 1362-3. 14. Black WA, Buskirk FV. Gentamicin blood agar as a generalpurpose selective medium. Appl Microbiol 1973;25:905-7. 15. Vaneechoutte M, Verschraegen G, Claeys G, Weise B, van den Abeele AM. Respiratory tract carrier rates of Moraxella (Branhamella) catarrhalis in adults and children and interpretation of the isolation of M. catarrhalis from sputum. J Clin Microbiol 1990;28:2674-80. 16. Kaulbach HC, White MV, Igarashi Y, Hahn BK, Kaliner MA. Estimation of nasal epithelial lining fluid using urea as a marker. J Allergy Clin Immunol 1993;92:457-65.