7 sran/8498-t06/402.0/0 Copia © 1965 by Wiliams e Wikins Co VoL 4, No.5 Printedin USA Efficacy of bone marrow, blood, stool and duodenal contents cultures for bacteriologic confirmation of typhoid fever in children CONSTANZA VALLENA‘ AND EDUARDO GOTU z0, MD ‘The relative efficacy of cultures made from duodenal contents (obtained by string capsules), bone marrow, blood and rectal swab was com- pared in 118 pediatrie patients, 2 to 13 years old with suspected typhoid fever. Only 47% of children 2 to 6 years old tolerated the string device, as compared with 89% in children 7 to 18 years old (P < 0.05) ‘The four culture techniques were performed and at least one was positive for Salmonella typhi in 43 patients. Bone marrow cultures were positive in 84% of the confirmed cases, a sensitivity significantly greater than for duo- denal contents (42%), blood (44%) and stool (65%) cultures. Higher recovery rates for blood cultures were found during the first week of illness than later (70 vs. 22%). Bone marrow cultures remain the most effec- tive method for the recovery of S. typhi. Stool cultures appear to be more effective in children than in adults. Duodenal contents cultures offer little advantage in young (2 to 6 years old) chil- dren. INTRODUCTION In endemic areas typhoid fever has a high incidence in children and young adults. In Lima it has been estimated that at least 25% of confirmed cases occur in children younger than 10 years old.’ Since very young children may have mild febrile but bacteremic illnesses not recognized as enteric fever,’ it is possible that the proportion of typhoid fever eases in children is even higher. Confirmation of typhoid fever requires isolation of Saimonella typhi during a typical illness, since serology has limited value in endemic areas.’ In adults culture Frou the Pediatrics Department (CV, HH) and the Instituto de Medicina ‘Tropical “Alexander von Humbolds,” Universidad Per tana Cayetano Heredia (2G), Lima, Per, andThe Jobne Hopkins University School of Medicine (BK) and University of Maryland School of Medicine (RB), Baltimore, MD. ‘Adeess for reprints: Eduardo Getuera, M.D. Universidad Per wana Cayetano Heredia, Apartado 5045, Lima, Peru 496, MD, HERMINIO HERNANDEZ, MD, BRADLEY KAY, MD, ROBERT BLACK, MD cultures of blood oF stool, resulting in isolation of S. typhi in 85 to 95% of bacteriologically confirmed cases." Likewise in children a culture of bone mar- row was previously demonstrated to be better than blood cultures, yielding the organism in 80 to 90% of cases. Aspiration of bone marrow for culture is recognized fs the optimal method for confirmation of typhoid fever and recovery of the organism to perform anti- biotic sensitivity studies," but because it requires trained personnel and sterile special equipment it is infrequently used in endemic areas. A search for a simpler and less costly culture technique led to the evaluation of culture of bile and duodenal contents obtained by a string which is swallowed in a gelatin capsule.*” In adults cultures of the string device have shown rates of recovery of S. typhi comparable to those of bone marrow aspirates."'"' Because such data were not available for children, we conducted a pro- spective comparison of cultures of duodenal contents, bone marrow, blood and stool in pediatric patients with suspected typhoid fever PATIENTS AND METHODS All children 2 to 13 years old with suspected typhoid fever at the Hospital General Base Cayetano Heredia between March and August, 1984, were studied. Cul- tures were obtained after consent of the parents and before antibiotic treatment was initiated. Patients were considered antibiotic-treated if they had received 48 hours or more of therapy in the week prior to admission, ‘A single 3-ml blood sample was inoculated into 10 ml of 10% Difeo OxGall” broth. A 0.5-ml bone marrow aspirate was also inoculated into 10 ml of 10% Ox- Gall®. A rectal swab was obtained by brushing the four sides of the rectal mucosa and then placed in selenite F enrichment broth, which was plated on SS agar after 24 and 72 hours of incubation. For culture of the duodenal contents, 70 cm of polyester yarn were packed in a gelatin capsule andVol. 4, No. 5 sterilized with ethylene oxide. After the patients fasted for 2 hours the capsule was ingested with sterile water, and the proximal end of the yarn was fastened with tape to the cheek. The yarn was removed 3 hours later and the distal 20 cm were placed in selenite F broth. ‘The color (bile-staining) of the distal 5 em of string ‘was recorded; the pH of the string was not tested. The selenite F broth was plated on SS agar after 24 and "72 hours of incubation and S. typhi was identified by standard methods (12). ‘The culture techniques were compared by the Coch- ran’s Q test (13), and other analyses were done by the chi square test. RESULTS Of 118 patients studied, 56 were male and 62 were female; 38 patients were 2 to 6 years old. S. typhi was isolated from 58 patients and Salmonella paratyphi from 4; other diagnoses were made for the remaining 56 patients. ‘Twenty-nine patients did not tolerate the string device; 13 could not swallow it and 16 vomited after swallowing. Children 7 to 13 years old tolerated the string better, with 71 of 80 completing the study, compared with only 18 of 38 children 2 to 6 years old (P < 0.05). Tolerance was not related to sex or ulti- mate diagnosis. ‘The four culture techniques were compared among the 43 patients in whom all techniques had been performed and at least one was positive for S. typhi. ‘Twelve patients were 2 to 6 years old and 31 were 7 to 13 years old; 21 were male and 22 were female. Fourteen had taken antibioties, 32 were untreated and three patients did not know what medicine they had taken. Twenty patients (56%) had had fever for 4 to 7 days, 18 (22%) for 8 to 14 days and 5 (12%) for 15 to 21 days. Five strings were not bile-stained (2 of 18 culture-positive and 3 of 25 culture-negative; P > 0.06). Bone marrow cultures were positive in 84% of the confirmed cases, significantly more often than the other techniques (Table 1). The string culture of duo- denal contents appeared to give a rate of recovery of S. typhi lower than that of stool and similar to that of blood culture. The relative efficacy of culture tech: niques was the same in patients who did or did not receive previous antibiotic therapy. The rates of re- covery of S. typhi from bone marrow, stool and duo- denal contents were similar in patients in the first week us. the second week of illness; however, higher recoveries were found in blood during the first. week of ilness (P <0.005). Bone marrow plus stool cultures together yielded 97% positivity. DISCUSSION ‘Typhoid fever in children presents a difficult diag- nostic problem because of atypical clinical aspects in children," false-positive agglutination reactions’ VALLENAS ET AL. 497 and the added difficulty of obtaining appropriate spee- mens for culture in children, In some previous studies in children with typhoid fever, blood cultures have yielded 80 to 90% positiv- ity." Other studies from Peru," * ®® Chile,?* Indonesia® and Mexico,’ which include adults and children, report 87 to 49% positivity. The positivity rate of 44% for blood cultures in this study is consist- ent with these reports, as is our finding of a higher rate early in the illness. Likewise our isolation rate of ‘84% from bone marrow aspirates and the observation that this high rate is not affected by the prior use of antibioties or the duration of disease confirms pre- vious diagnostic studies in adults and children typhoid fover.* In other studies stool cultures have often yielded low rates of positivity, ranging from 19 to 44%." ¥-2.21 Tn contrast, in our study stool cultures were considerably more effective, being positive in 65% overall and in 55 to 74% regardless of the length of illness or previous antibiotic therapy. This rate of positivity for stool cultures is higher than that ob- served in adult patients in our hospital during the same time period (13 of 48; 27%). Using the same bacteriologic techniques and considering only patients without previous antibiotic treatment, we found that stool eultures yielded 73% positivity in children and 35% positivity in adults (P < 0.005). Thus in this patient population there appears to be a more numer- us or frequent fecal excretion of S. typhi in children than in adults, but it is possible that the technique for stool collection was better in children (rectal swabbing in four sides) than in adults (simple swabbing or direct stool sampling). Cultures of duodenal contents were positive in only 42% of pediatric patients with confirmed typhoid fever and substantially less than in adult patients." Bile staining is a suitable marker of duodenal location. TABLE 1 requency of isolation of Salmonell typhi from cultures of various sites in 43 patients with typhoid fever by previous antibiotic therapy ind duration of lines Positive Cnlsars in Vor Sites Da Characteristics BOM Stoo) Blood ‘dena sents Ne @ No ® No No Previous antibiotic 2% 857 6 6 ast 19 1 410 8B Dunatlon a nese ‘rane (Wea) 16 89 1 S14 T0(A) 7-5 Siigevon) 2081 aS te) ab ) 8 65 19 Ms teat; ©, P= 0108, Cochran's Q498 ‘The possibility of having false-negative values was quite low since we found that only 3 of 25 culture- negative strings were not bile-stained. In addition acceptance of the procedure was not good in children less than 6 years old, of whom more than half could not tolerate the string device. Thus the culture of duodenal contents seems to offer little advantage in the youngest children and could be useful in older children only if the rates of positivity were improved. It is possible that the strings require more time to reach the duodenum in children and that the 3-hour ‘period for collection of the culture was too brief. We are currently evaluating a 4-hour culture period and are testing the pH of the distal segments of the string after withdrawal to demonstrate that it has entered the duodenum, At this point bone marrow cultures remain the most effective method for the acteriologic confirmation of thyphoid fever. ACKNOWLEDGMENTS “This study was supported by grants from the National Institutes (of Health (AI 20150-01) and the Nestle Coordination Center for Nunition ‘We thank Drs. Miguel Campos, Claudio Lanate, Humberto Guerra, Emilio Crosby and Augusto Vi for their advice and ertique of the manuscript. ‘We also thank the staff of the Microbiology Laboratory of the Universidad Peruana Cayetano Heredia for thet assistance and the personnel of the Pediatris Department of the Hospital General Base Cayetano Heredia for relerrl of patients and continuous help with tis study. REFERENCES 1, GotuzzoB, Gueree J, Crosby B, otal: valuacion dels proce- dlimientos diagnosticos de febretfodea. Acta Med Peruana saeig, 78 2. Ferreccio Levine M, Manterola A, ota: Benign bactreaia caused by Solmaneta phi and paraph n children younger than 2 year. J Pediatr 106880-807, 1008 8. Levine Mf, Grados , Gilman Reta: Diagnostic valve of che Wid esti areas endemic fo iyphoid fever Arn J Trop Med Hyg 2:76-800, 1078 4 Guera J, GovurzeB, Cosby B, eta: Diagnostic value of bone ‘TYPHOID FEVER DIAGNOSIS IN CHILDREN September, 1985 marrow eulture in typhoid fever. Trans & Soe Trop Med Hyg 731680-653, 1979 Gotuzeo By Delgedo E, Maguina ©, et al: Tratamiento de la Fiebre tifoides con Bacampiilin. Abstracts of the Second Pon- ‘american Congress of Infectoogy, Buenos Aires, Argentina, 1984 Hoffman §, Punjabi N, Rockhill R, et al: Duodenal string capsule eulture compared with bone marrow, blood and reel swab cultures for diagnosing typhoid and patatyphoid fever. J Info Dig 140167161, 1984 Gilman R, Torminel Mt, Levine M, et al: Relative efficacy of blood, urine, rectal swab, bone marrow and rose spot cultures for recovery of S. phi in typhoid fever. Lancet 11211-1213, 197 (Chang 8, Hermandex H, Vi A, etal: Hemocultivo y mielocultivo tio con fiebretifoides, ol Med Hosp Infant Mex 2614— 6, 1982 Gilman R, Hornick R: Duodenal isolation of Salmonela typhi by string cape in acute typhoid fever. J Clin Microbiol 8456, 1976 Benavente L, Gotuzzo I, Guerra J etal Diagnosis of Solmo- nella typhi by ealture of dvndenal string eapsule letter). V Engl J Med 80:54, 1981 Benavente L, Gotuzz0 B, Guerra J, eal: Diagnose of typhoid fever using a string caps devie, Trans R Soe Trop Med Hy ‘304-106, 1984 awards PR, Baing WH: Identification of Enterobactriacene, “Minneapolis, Burgess Publishing Company, 1972, p 116 Pieiss J: The comparison of m-matched samples. Statistica! Methods or Rates and Proportions. Second Raition. New York dJohn Wiley & Sons, 1981, pp 120-133, Christie AB: Typhoid and peratyphoid fevers. Infectious Dis teases: Bpidemialoy ond Clinical Practice. Fdinbuegh, Churchill Livingstone, Lid, 1974, pp 85-121 Mulligan T6: Typhoid fever in young children, Br Med J 4:665— 667, 1971 Khuri-Bulos N: Enteric fevers in ehildren, Clin Pediats20:648- 452, 1981 ‘Jobson A, Aderele W: Enterie fever in childhood. J Trop Medd Hyg 8428-85, 1881 Heredia A, Trevino J: Estudio clinco y de lsborstorio de 437 ninos con febre tfoides. Bol Med Hoxp Infant Mex 17:728-742, 1960 2 Scack L, Pino M, Wiederhold A: Bl mislocultivo en el diag nostico do fibre tifeidea yparetifiden. Rev Chil Pediatr 7218-220, 1966 Kurnate J, Gutierrez G: Fiebre tifoidea. Manual de Infectologia. torial Medica del Hospital Infantil de Mexico, 1977, pp 11-21 Colon Ay, Gross D, Tamer Mf: Typhoid fever in children. Pdi tres 56:606-608, 1875,