Elsevier

The Lancet

Volume 365, Issue 9454, 8–14 January 2005, Pages 130-134
The Lancet

Articles
Induced sputum versus gastric lavage for microbiological confirmation of pulmonary tuberculosis in infants and young children: a prospective study

https://doi.org/10.1016/S0140-6736(05)17702-2Get rights and content

Summary

Background

For microbiological confirmation of diagnosis of pulmonary tuberculosis in young children, sequential gastric lavages are recommended; sputum induction has not been regarded as feasible or useful. We aimed to compare the yield of Mycobacterium tuberculosis from repeated induced sputum with that from gastric lavage in young children from an area with a high rate of HIV and tuberculosis.

Methods

We studied 250 children aged 1 month to 5 years who were admitted for suspected pulmonary tuberculosis in Cape Town, South Africa. Sputum induction and gastric lavage were done on three consecutive days according to a standard procedure. Specimens were stained for acid-fast bacilli; each sample was cultured singly for M tuberculosis.

Findings

Median age of children was 13 months (IQR 6–24). A positive smear or culture for M tuberculosis was obtained from 62 (25%) children; of these, 58 (94%) were positive by culture, whereas almost half (29 [47%]) were smear positive. Samples from induced sputum and gastric lavage were positive in 54 (87%) and 40 (65%) children, respectively (difference in yield 5·6% [1·4–9·8%], p=0·018). The yield from one sample from induced sputum was similar to that from three gastric lavages (p=1·0). Microbiological yield did not differ between HIV-infected and HIV-uninfected children (p=0·17, odds ratio 0·7 [95% CI 0·3–1·3]). All sputum induction procedures were well tolerated; minor side-effects were increased coughing, epistaxis, vomiting, or wheezing.

Interpretation

Sputum induction is safe and useful for microbiological confirmation of tuberculosis in young children. This technique is preferable to gastric lavage for diagnosis of pulmonary tuberculosis in both HIV-infected and HIV-uninfected infants and children.

Introduction

Diagnosis of pulmonary tuberculosis is difficult in infants and young children in whom clinical and radiological signs can be non-specific and variable.1 This difficulty has been compounded by the HIV epidemic, because other HIV-associated lung diseases mimic the clinical and radiological picture of tuberculosis, the development of anergy has reduced the sensitivity of skin testing for tuberculin, and clinical scoring systems have not been developed specifically for HIV-infected children.2, 3

Microbiological confirmation of tuberculosis is desirable for definitive diagnosis, for best use of antituberculous medication, and for epidemiological tracing of isolates. In infants and young children, culture confirmation has relied on specimens from sequential gastric lavages.3 Although sputum induction has been successfully used in adults,4, 5, 6 this technique is not regarded as feasible in young children since they swallow their sputum and do not expectorate. Limitations of gastric lavage, however, include the need for an overnight fast, repeated specimens, and admission of children. Moreover, the procedure is time consuming and unpleasant for both child and health worker. By contrast, sputum induction is less invasive than lavage, quicker to do, and can be done in resource-poor settings or outpatients.

Studies comparing gastric lavage with sputum induction in adults with suspected tuberculosis have reported that the diagnostic yield from sputum is higher.7, 8, 9, 10 However, few studies of sputum induction in young children are available. A study of 30 Malawian children with suspected tuberculosis reported that diagnosis could be confirmed by staining or culture of induced sputum in eight individuals (28%); however, most children were older than 5 years, and the yield from gastric lavage was not measured.11 In a previous study of young children (median age 9 months) admitted for acute pneumonia,12 we reported that sputum induction could be safely and effectively done. We also reported that in 16 children with tuberculosis confirmed by culture, the yield from induced sputum was equivalent to that from gastric lavage.13 However, few children had tuberculosis, only one sample from induced sputum was compared with one to three from lavage, and clinical suspicion of tuberculosis was low since children were admitted for acute pneumonia.13 The aim of the present study was to compare repeated specimens from induced sputum with repeated gastric lavages for yield of Mycobacterium tuberculosis in infants and young children with suspected pulmonary tuberculosis.

Section snippets

Patients

The 2-year study was done from June, 2000, to June, 2002, in the general paediatric wards of two hospitals in Cape Town, South Africa—Red Cross War Memorial Children's Hospital and Somerset Hospital. Children were enrolled if they had been admitted for suspected pulmonary tuberculosis on the basis of a chronic cough (more than 28 days) and one of the following criteria: household contact known to be infected with tuberculosis within the previous 3 months; loss of weight or failure to gain

Results

250 children underwent sputum induction and gastric lavage; 141 (56%) were male and the median age was 13 months (IQR 6–24). Baseline median respiratory rate of children was 56 (40–64) breaths per minute, and median arterial oxygen saturation was 96% (95–98%). 68 (27%) children were receiving supplemental oxygen at the time of sputum induction; 65 via nasal prongs or cannulae and three via headbox oxygen. 30 children (12%) were known to be HIV-infected at enrolment; of the children whose HIV

Discussion

We have shown that the diagnostic yield from sputum induction was better than that from gastric lavage in infants and young children admitted for suspected pulmonary tuberculosis. One sample from induced sputum yielded twice the number of positive cases than did a single specimen from gastric lavage. The yield from one sample from induced sputum was equivalent to that from three sequential gastric lavages, the recommended clinical practice for microbiological confirmation in infants and young

References (24)

  • DT Carr et al.

    A comparison of cultures of induced sputum and gastric washings in the diagnosis of tuberculosis

    Mayo Clin Proc

    (1967)
  • H Kawada et al.

    The usefulness of induced sputum in the diagnosis of pulmonary tuberculosis

    Kekkaku

    (1996)
  • Cited by (418)

    View all citing articles on Scopus
    View full text