"Drug resistance impairs response to treatment, extends the period of infectivity, and is associated with high mortality", says Daniel Orozco of MSF, which supports 22 laboratories in Uzbekistan.

MSF is seeing rates of multidrug resistance as high as 13% among new cases presenting to the clinics, and plans to start a pilot programme to treat such patients this year.

WHO has declared the ongoing tuberculosis epidemic a global emergency, one that is expected to kill 30 million people in the next decade, mostly in the developing world. The global expansion of the WHO-led directly observed therapy, shortcourse (DOTS) strategy for the treatment of tuberculosis is a significant advancement in addressing this disease, but WHO has long since advocated that tuberculosis control lies not only in effective treatment delivery, but also in availability of laboratory support for early case detection and diagnosis.

Yet, enter a tuberculosis hospital laboratory in Uzbekistan, central Asia, a country in the throes of a huge tuberculosis epidemic but whose health services barely function, and the needs remain great.

The laboratories are basic and unsafe by international standards; supplies are sporadic, equipment broken, and quality assurance badly monitored. This situation is probably generalisable to most other resource-poor countries.

"Well-functioning laboratories in disease-endemic countries are critical to ensuring high-quality examinations, consistent reporting, and detecting incident tuberculosis cases", says Mark Perkins (UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases). "Yet this is hard to achieve with few resources".

Poorly functioning laboratories miss many tuberculosis cases, the detection of which is already difficult given the solitary focus of the DOTS strategy sputum smear microscopy for case detection.

Last year, although 16 of 22 countries with the highest burden of tuberculosis reported treatment success rates of more than 70%, only four of them had overall smearpositive detection rates of more than 60%. Those cases that are detected have often been symptomatic for months, despite previous visits to health clinics.

"A low rate of case detection and long delays in reaching a diagnosis translate into large numbers of source cases persisting in the community, stoking the epidemic. The poor state of laboratories leads, in turn, to poor performance, perpetuating a vicious cycle of laboratory mediocrity reinforcing clinical irrelevance", says Perkins. The neglect of such services is most often a financial resource issue, but not exclusively. The diagnostic tool of sputum smear microscopy itself is at fault, for it can only detect a minority of tuberculosis patients, even when done well.

Apathy on the part of the international community to promote long-term investment into the development of more sensitive and rapid diagnostic tests, and sustainable laboratory systems, does not help. What resources there are go largely to drug treatment to increase cure rates. "Part of the reason lies in the age of the technology involved", says Perkins. "If the main tool for diagnosing tuberculosis was only a few years old - like HIV tests - there would be substantially more interest from donors and host governments in ensuring quality services." For a future with sound measures for tuberculosis control, DOTS is clearly not going to be enough.

Recognition of disease transmission by smear-negative cases, as well as the threat of drug-resistance, now demands that culture and drug susceptibility testing (DST) should be the standard of tuberculosis care globally.

Yet attempts at this will have huge resource implications for laboratory services in resource-poor countries. "Drug resistance impairs response to treatment, extends the period of infectivity, and is associated with high mortality", says Daniel Orozco of Médecins Sans Frontières (MSF), which supports 22 laboratories in Uzbekistan. MSF is seeing rates of multidrug resistance as high as 13% among new cases presenting to the clinics, and plans to start a pilot programme to treat such patients this year.

"But to uncover the extent of the problem, MSF flew all samples to a WHO-accredited supranational laboratory (SRL) in Germany that was willing to fund transport and processing. The survey cost US$39,000 to do, which makes such work totally unrealistic for resource-poor countries, thus masks the extent of resistance globally."

Although there is a network of 20 nationally financed WHO-accredited SRLs, whose remit includes supporting DST surveys, this status does not oblige these laboratories to carry out any such function and some do effectively nothing. "We all joined this laboratory network because we felt there was a need to offer expertise and support to resource-poor countries, and it is part of our remit", says Francis Drobniewski (Health Protection Agency, London, UK).

"The reality is that the WHO doesn't provide anything to this network - no funds and no support. We would like to address the issue of drug resistance, but we can't do this work for free." WHO say they are working to address some of these issues.

The success of future global efforts to control tuberculosis will depend on the ability to detect patients early and interrupt transmission cycles through strengthened laboratory networks and developing reliable, cheap diagnostics that work beyond the capacities of sputum smear microscopy. In the face of drug-resistant disease, this is more crucial than ever. It remains to be seen whether sufficient will can be mustered internationally to push these issues further up the agenda. Sally Hargreaves