Evaluation of Microscopic Observation Drug Susceptibility (MODS) and the string test for rapid diagnosis of pulmonary tuberculosis in HIV/AIDS patients in Bolivia

Table 3 String Test sub-analysis: Patient characteristics and comparison of diagnostic methods in sputum productive, sputum scarce and sputum absent patient subgroups

Patient Characteristics (n = 102)	Sputum-productive (n = 70)		Sputum-scarce (n = 22)		Sputum-absent (n = 10)
Median age (years)	32		30.5		36
CD4 count	122 (IQR 45–271)		55 (IQR 23–86)		221 (IQR 133–486)
Rate of TB+	28/70 (40 %)		11/22 (50 %)		4/10 (40 %)
Mortality of TB+	10/28 (35.7 %)		8/11 (72.7 %)		0/4 (0 %)
Test	+Mtb (n = 28)	Diagnostic yield (95 % CI)	+Mtb (n = 11)	Diagnostic yield (95 % CI)	+Mtb (n = 4)	Diagnostic yield
String test MODS	18	64.2 % (44.7-81.4)	9	81.8 % (48.9-97.7)	4	^a
Sputum MODS	27	96.4 % (81.7-99.9)	11	100 % (71.5-100)	^a	^a
Sputum LJ	21	75.0 % (55.1-89.3)	10	90.9 % (58.7-99.8)	^a	^a
Sputum ZN	17	60.7 % (40.6-78.5)	7	63.6 % (30.8-89.1)	^a	^a

TB positive was defined as any patient with sputum or gastric sample with a positive culture result from sputum or string test by LJ or MODS. This composite positivity was used as the reference standard to determine diagnostic yield. The patients that submitted a string test (n = 102) were divided into three groups based on degree of sputum productivity. Sputum-productive patients (n = 70, 68.6 %) were able to produce spontaneous sputum. Sputum-scarce patients (n = 22, 21.5 %) required sputum induction to produce sample. Sputum-absent patients (n = 10, 9.8 %) were unable to produce sputum despite induction and only submitted a string sample. Of the sputum absent patients, 4 patients were positive for Mtb by string test, but there are no samples for comparison since they were unable to produce sputum. The string test samples were not neutralized prior to culture and this may have affected diagnostic yield
n number of patients in group; IQR interquartile range; CI confidence interval
^a We are unable to provide these numbers as sputum absent patients did not submit sputum for comparison

ISSN: 1471-2334