An assessment of nucleic acid amplification testing for active mycobacterial infection

Component	A Excellent	B Good	C Satisfactory
D Poor
Evidence-base^a				Level IV studies, or level I to III studies/SRs with a high risk of bias
Consistency		Most studies consistent and inconsistency may be explained
Clinical impact				Slight or restricted
Generalisability		Population(s) studied in the body of evidence are similar to target population
Applicability		Applicable to Australian healthcare context with few caveats

SR = systematic review

^a Level of evidence determined from the NHMRC evidence hierarchy (see Table ).

Source: Adapted from NHMRC (2009)

Ponticiello et al. (2001) presented the results of a poor-quality prospective cohort study of TB patients and their close contacts to investigate sociodemographic and clinical risk factors for transmission of TB, including delay in diagnosis (Table ). The study was conducted at a TB referral centre in Italy, and included patients with newly diagnosed pulmonary TB during the period January 1997 and December 1998, and their close contacts. Close contacts included the patient’s household and all other persons sharing the same indoor environment with the patient for prolonged periods. The applicability of this study is limited by the fact that the effect of extended delay to diagnosis was compared with a diagnostic delay of up to 1 month; the risk of transmission to contacts when treatment was initialised immediately upon development of symptoms was not reported. In addition, delay in diagnosis was defined as the period from onset of any TB symptoms to the diagnosis of TB, which includes both the patients’ delay in seeking medical attention (presumably as reported by the patient) as well as health-system delays in diagnosis and treatment initiation.

All the patients included in the study had AFB microscopy of sputum or bronchial specimens. The average delay in diagnosis of TB was 2.25 ± 1 month. In a multivariate logistic regression model, delay time to diagnosis was the only factor that remained significantly and independently associated with an increased risk of infection among contacts, as determined by positivity to the tuberculin skin test (p<0.0002). For patients with a diagnostic delay of 1.5 months the adjusted odds ratio for contacts infected / not infected, compared with a diagnostic delay of less than 1 month, was 4.2 (95%CI 1.3, 13.7), while with a diagnostic delay of 2 months this increased to 6.1 (95%CI 1.9, 19.6).

Golub et al. (2006) performed a similar prospective cohort study of poor quality to determine the association between total treatment delay and TB transmission in patients with verified pulmonary TB reporting to the Maryland Department of Health and Mental Hygiene in the USA between June 2000 and November 2001 (Table ). Total treatment delay was defined as the interval from first TB symptoms to initiation of treatment for TB. The median total treatment delay for US-born patients was 99 days, with 67% (36/54) of patients having delays ≥ 60 days. The probability of having infected contacts (tuberculin skin test positive) increased with longer delay in diagnosis; 38% of contacts of patients with a delay of ≥ 60 days had positive tuberculin skin tests compared with 25% of contacts with < 60 days’ delay (p=0.05). Although these results are not surprising, they reinforce the belief that quicker diagnosis of TB is of great benefit in reducing its spread to close contacts of infected individuals.

van der Oest, Kelly & Hood (2004) performed a retrospective study of poor quality based on notified cases of TB among residents of the Waikato Health District, located in the North Island of New Zealand, from January 1992 to December 2001 (Table ). Although the outcome measure used in the study—favourable treatment outcome (i.e. cure or treatment completion)—was not specified in the PICO, due to the paucity of relevant evidence on the health impact of treatment delay on the patient, this study was included. The definition of diagnostic delay was the time between development of symptoms and diagnosis. While 84.7% of patients were reported as having delayed diagnosis of over 4 weeks, the mean and standard deviation of the time to diagnosis were not provided. Of those patients for whom data were available, 79% successfully completed treatment. The results of a logistic regression model indicated that time between development of symptoms and diagnosis was not significantly associated with a favourable treatment outcome (OR=1.02; 95%CI 0.99, 1.04; p=0.15). As ‘favourable treatment outcome’ was poorly defined in this study, this result may simply reflect that the treatment completion rate, which may be influenced by many factors, appears to be unrelated to any treatment delays.

Table Summary of studies assessing the health impact of early versus delayed treatment of TB on the individual and their contacts