Sputum smear microscopy remains the cornerstone of tuberculosis (TB) diagnosis for the vast majority of patients with TB around the globe. The advantages of near-patient availability in peripheral healthcare facilities and a clinically useful turnaround time of a day or two mitigate to some extent the relatively poor sensitivity of smear microscopy, which results in around half of patients with active pulmonary TB (i.e., smear-negative but culture-positive) being missed (1). Efforts to improve analytical sensitivity have intensified in recent years, and notable advances have included the demonstrated effectiveness of simple instruction strategies to improve sample quality (2), the clear benefit afforded by almost all concentration methods (3) (though these incur financial, manpower, and time costs), and the transformation of fluorescence microscopy from the preserve of wealthy laboratories armed with a steady supply of expensive replacement mercury bulbs to a widely applicable tool through the availability of relatively inexpensive and durable LED-based technology (4). The clearly decreasing incremental returns with successive smears (5) has led the World Health Organization (WHO) to revise recommendations from the conventional reliance upon three diagnostic sputum smears down to two (6), at a stroke reducing microscopy laboratory workload and thereby (it is hoped) allowing quality to improve.

Thus the greatly increased potential availability of low-cost, more efficient smear microscopy at the peripheral reaches of the healthcare system would seem to be a big stride forward. But tuberculosis is primarily a disease of the poor, who are then further impoverished by it; seminal work from Malawi has taught us that even a free TB test can bankrupt the extremely poor (7). When patient and household direct and opportunity costs were counted, the poor still spent 248% of monthly household income accessing free TB diagnostic services within 6 km of their home. Against this background and with an eye on reducing the time to complete smear microscopy workup, there emerged the concept of front-loaded smear microscopy, wherein the results obtained from two samples taken on the same day at least 1 hour apart were shown to be equivalent to two samples on different days (8). This innovative approach was endorsed by WHO earlier this year (9), and holds great promise for streamlining the patient diagnostic journey and health system efficiency.

In this issue of the Journal (pp. 547–551), Cattamanchi and colleagues report results from a study undertaken in the high-TB and high-HIV burden setting of urban Uganda that could further transform the global approach to sputum smear microscopy (10). As so often happens in the field of tuberculosis, the old literature reveals important opportunities to relearn the lessons our predecessors so clearly described. Many of the approaches described above aim to overcome the fact that the number of bacilli present in sputum samples is low (and close to the limit of microscopic detection) by increasing the volume of sputum examined, either through concentration (thus a greater proportion of the bacilli in the sample appears on the smear) or through examination of multiple samples. Cattamanchi and coworkers hypothesized that performing a second smear from a single unconcentrated sputum sample would be equally as effective as performing a second smear from a separate, second sputum sample; they further explored whether this hypothesis would hold for both Ziehl-Neelsen (ZN) and fluorescence microscopy. The data clearly demonstrate that for patient diagnosis, fluorescence microscopy outperformed ZN light microscopy regardless of the sputum sampling strategy, and that two smears performed on the first sample yielded the same performance as two smears each performed on different samples taken on separate days—the incremental gain from the smear of the second (early morning) sample was no greater than simply performing a second smear on the first sample, for both ZN and fluorescence microscopy.

In an arena sometimes overrun with hype, this observation looks to have the genuine potential to be paradigm changing. A single patient visit (or perhaps two if the return for results is needed) is obviously desirable for all the same reasons that make front-loaded microscopy so attractive. The simplified logistics of only having to collect half as many diagnostic samples could importantly reduce workload and cost. What, then, are we waiting for? As always, there are some interesting questions thrown up and a few caveats to consider. What would be the effect of concentrating the single sputum prior to the two smears? How would two smears on one sample compare with one or two smears of the same sample after concentration? It might be that a small number of additional cases could be detected, but would that justify all the extra concentration work? How would improved sputum sample quality (through clear instructions [2]) play out when one rather than two samples were being collected? The definition of sputum smear-positivity used in this work was greater than or equal to 1 acid fast bacilli (AFB) per slide, as recently accepted by WHO “where a functional external quality assurance for smear microscopy is in place”—though the older definitions, requiring (for ZN smears) greater than or equal to 10 AFB per 100 high-power fields for a smear to be called AFB 1+, are still widely used. Given that uptake of the “1 AFB per slide = positive” message may be slow and reluctant, it will be important to understand how the different positivity threshold criteria affect performance of this approach. An unexplained oddity in this study is that the incremental yield from the second smear, whether from the same or a second sample, was considerably lower than would have been expected—4% and 5% for ZN and 3% and 6% for fluorescence microscopy, compared with an average of 11.9% reported in a systematic review (5). Whether this reflects higher than normal sensitivity of the first smear (perhaps due to the 1 AFB threshold) and therefore lesser opportunity for the second smear to have an impact, or generally lower sensitivity of the smears, is not clear.

It will be important to repeat this work in a range of other settings before reaching the conclusion that this is how smear microscopy should now be done. For example, it would be critical to determine whether these findings from a research study undertaken in a busy urban African hospital can be reproduced in an operational environment in rural health center microscopy units in Asia, Latin America, and Africa. If they can, then we might really have something that changes the rules of the game.

References

Section:

ChooseTop of pageReferences <<CITING ARTICLES