First to the draw but with another question and not an answer. I would be grateful for pointers on how different "strains" in the MTB complex are classified- i know africanum, bovis and other types exist. The second question is the more fundamental question of what level of genetic variability constitutes a strain. I know from biology that it is feathers rather than wings which define birds.

In an attempt at an answer, I would cite host factors as being the main reason why some subjects do not get active disease - from my knowlegde of the clinical presentations of mycobacterial disease. The case of mycobacterium leprae and susceptible hosts like the armadillo etc

Thanks for a very important question, but I think it throws up further questions which will be difficult to answer.

There are several excellent review articles that address the conditions that are associated with the reactivation of 'latent' Mtb. The most recent coming from the Mayo Clinic in July of 2013. In summary, Mtb reactivates when there is a change in the host's immune status induced by chemotherapy, malnutrition, degradation of the immune system due to aging, other systemic diseases (diabetes, hematologic and some solid tumor malignancies), long-term steroid use, end-stage renal disease, etc. Reactivation is a variance of the host's ability to control the infection and not something related to the organisms pathogenicity.

Hi Leonard. Both the host and the pathogen might play a role. I have experience only in animal hosts of the MTC, and there, strain differences are not very evident, while host factors have a clearer influence.

Hi Leonard:

Reactivation of MTB is commonly associated to an immunodeficiency. With the advent of HIV/AIDS, this change from latent to active TB has been observed among people living with HIV in several resource-limited settings. Of note, TB prevalence is as higher as 40-50% in this target population.

Certain strains do progress from latent Mtb infection to active disease more quickly than others. For example in this study of case contacts in The Gambia M. tb complex organisms progressed to active disease more quickly than M. africanum (HR 3.1) .(1) However other host factors also play a role in progression, and may have a greater effect than strain. This includes HIV infection (8-40x risk of progression), other forms of immune suppression such as that caused by drugs (corticosteriods, TNF antagonists), diabetes and malnutrition. I do not find the 'parasite never kills its host' axiom to be useful in infectious diseases, Mtb alone kills over 1.2 million people a year! Latency may have some advantages to Mtb but transmission is essential to Mtb's continue survival and that requires active pulmonary disease. Those strains that most readily reactivate and induce a greater tissue necrosis are thought to have an advantage. (2). Necrosis favours tissue destruction (cavitation). heavy excretion of bacilli in sputum and therefore maximizes transmission.

(1)de Jong BC, Hill PC, Aiken A, Awine T, Antonio M, Adetifa IM, et al. Progression to active tuberculosis, but not transmission, varies by Mycobacterium tuberculosis lineage in The Gambia. J. Infect. Dis. 2008 Oct 1;198(7):1037–43.

2. Parwati I, van Crevel R, van Soolingen D. Possible underlying mechanisms for successful emergence of the Mycobacterium tuberculosis Beijing genotype strains. The Lancet Infectious Diseases. 2010 Feb;10(2):103–11.

The common terms have been latent, non-replicative, or dormant , which are inadequate. But I like the term you use benign. Is there a benign life cycle for TB? It's an interesting question. Having said that, TB infections can cause cavities without major symptoms. Cavities are self-healing and can continue for life until something happens and the clinical manifestation of cavitary TB develops. However, active TB lesions can be misdiagnosed as other lung diseases. Look up studies by Robert Hunter in this regard. Some studies in primates support the notion of a disease process that festers subclinically. Recent work on autophagy may provide clues to your question in the future.

This is an interesting question and I agree with other commentators that both immune mechanisms and species as well as strain/lineage are important in progression of disease. Pulmonary M.avium-intracellulare is usually related to compromised immunity, either local or systemic and seldom otherwise disseminates or progresses locally whilst in youngsters it appears mainly in tuberculous lymphadenopathy. Further information on this in M tuberculosis comes from the timecourse of extended outbreaks in families or communities that are in low prevalence areas and where outbreaks persist despite active case-finding. I have seen a Beijing outbreak persist in a rural enclave for 17 years despite active case finding including for tuberculin or lymphocyte reactors as each case was diagnosed. It seems to em that infectivity and latency ability are not the same thing but is is unwise to assume that latency implies benign strains as in time latency may well break down leading to clinical infection. It would be interesting to know about the human genotypes and strains of intrafamilial spread and whether progression or latency in a family member is predictable from knowledge of both. does anyone know if this has been done?

Agree that this is an interesting question and I agree with what has already been presented. Having carried out many contact investigations I would add that 'dose' is also a factor. TB cases with advanced pulmonary or laryngeal disease are more likely to infect more people....and more people explosed to highly infectious cases (representing high bacterial burden) are likely to develop TB disease. This does not contradict all the factors already mentioned, it is simply another factor to consider.

That is I think is true. We found in mice that 100 cfu will kill GM-CSF knockout mice by 4 weeks, but it takes 6 months with 10 cfu. So there appears to be a "Goldilocks" relationship between the strength of the immune system and the strength (in this case possibly strength in numbers) that keeps a fantastically active pathogen in check.

Most adult TB occur by reactivavión, it must be assumed that the immunity of the host is able to control the development of the disease. At low immune proliferation occur and the consequent development of tuberculosis. Is the case with most pulmonary TB, which occur due to reactivation of the primary complex. The disease is whether the balance between the ability of the host to defend itself, versus the infectious agent of developing the disease, so there will be more virulent germs and men will be guests to defend themselves.

with respect to Mycobacterium tuberculosis infection:Lot depends on the immune defense mechanism of the host (human) rather than the bacterium itself

Leonard,

I will give an indirect answer to your question. In India BCG vaccination is mandatory for all babies...however if due to any reason it cannot be given upto the age of 3 years, there is no point in administering it anymore. Reason? By then the majority of the Indian population has already come in contact with the bacterium and the necessary protective steps have already been taken by the immune system. As others have said developing the disease is dependent to a large extent on the host immune system.

As regards the term "Benign strain" that you have used, i have never really thought about it that way...very interesting.

There is no benign strain of M TB.But the bacillary load counts a lot given a particular level of immune status.BCG immunization status ,immune status of host plays a major role in developing countries.Infection turns to a disease when the host can not contain the organisms due to the depleting immune state or immunosuppressing events ie drugs,malignacy etc.Virulence of a strain may play a role but probably is immaterial in converting an infection to disease.

It has been observed that primary TB progresses to post primary TB in 5-10% of infected persons . The non progression is due to immunological mechanisms .The cell mediated immunity by T cells prevents progression of TB , while delayed type of hypersensitivity is responsible for progression . HIV disease highlights the role of the immune system in TB by destroying the T cells & progression of primary TB to post primary TB occurs in 80% of patients . The immunological mechanism probably determine the development of active TB disease in patients with TB infection . The clinical features of post primary TB in adults is a bilateral upper lobe fibrocavitary disease , while in advanced HIV disease it is lower lobe interstitial type of disease , similar to miliary TB , highlighting the role of immune mechanisms .

There is no question that the immune system plays a role in development of active TB. However, the question addresses a simple point: are the strains that are isolated from granulomas of patients at autopsy (shall we say: above 60 yrs of age) that have been free of any clinical symptoms (sub-clinical) significantly different from those isolated from active TB cases? TB associated deaths, in my opinion are opportunistic situations independent of the infection; namely, due to factors other than the infection itself (example: nutrition; stress, strife; systemic infections; etc). In other words, the existence of a non-virulent strain of Mtb that is associated with death is coincidental rather than causative. The answer to this question lies in the isolation and the characterization of that Mtb strain obtained from autopsy cases where the infection is noted as a coincidental finding . This study has as of yet not been done and may be of interest for the understanding of tuberculosis. Perhaps a study conducted in the US involving autopsies of immigrant subjects whose origins are from areas of the globe that have high frequencies of pulmonary tuberculosis and associated deaths. The change of environment may make little difference with respect to virulent strains vs benign strains. Just a thought.

Dr Leonard Amaral suggestion to do autopsy in immigrants from areas with high incidence with TB is interesting . I would suggest that it includes patients who have severe HIV disease & subjects with other immunodeficiency states due to various factors, who did not develop the clinical TB disease & died due to other opportunistic infections. This would highlight that the TB organism isolated did not cause the clinical TB disease in this group , suggesting a benign strain . Immigrant subjects with good health , who die of other causes may not have TB disease & benign strain cannot be excluded , even if TB bacilli is isolated . TB is common in developing countries & even in these areas , lower socioeconomic groups are most vulnerable due to multiple factors impacting their immune status .

I just checked if there were differences in MTC type frequencies between healthy adult wild boar with low TB lesion scores and individuals of any age with high lesion scores: there are! It would be fantastic to compare this with similar datasets from other species, human cases included.

The answer is mostly in the host immune response. Especially the activity of the IL-12/interferon-gamma axis. People with deficiency of this pathway are at increased risk of TB, particularly disseminated infection. It is this and other, yet to be identified, pathways that are most likely the reason why some people get sick from TB and others don't

Although there is no question that the immune system has an impact on who is to progress to active disease, the question remains unanswered. Are strains that are associated with granuloma noted at autopsy for whom the subject has not progressed to active disease status different from those that are associated with active disease.

If there is a lesson to be learnt from parasitology, it is this: Parasites do not kill their hosts. Death is incidental as a consequence of external factors, for if they were to kill their hosts, they to would perish, etc. Any interest in the obvious study that needs to be done?

Are there any studies that tested mycobacteria from autopsy in model hosts? Mice, guinea pigs etc. A parasite or a type of an "adopted" commensal? Does the host derive any benefit from being colonized by Mtb? Mycolic acid is a potent adjuvant. It may help keep lung mucosal immunity at the right state of activation until the rules of this treaty are broken.

Latent TB occurs in 90% of cases , because the human immune system can contain them inside a granuloma , but does not destroy them . There is no need to treat this group . In 10% , the immune system is not able to contain them & in the ensuing battle , the lung is damaged & the TB organism can spread to other human hosts .This needs Anti Tuberculosis drugs to destroy the bacteria . But MDR MTB & XDR TB have revealed that MTB can overcome the effect of drugs . This is in immunocompetent patients . The HIV epidemic has revealed that MTB can cause a disseminated disease with fatal consequences , as the human immune system is destroyed . The MTB is a virulent organIsm & latent TB should not be considered due to a benign organism , but an effective immune system . The MTB can survive this hostile environment for years & can reactivate at any time , by an error of the human immune system .

The recent debate on antibiotic resistance has highlighted the concern that , if new drugs are not discovered , we may go back to the pre antibiotic era , with disastrous consequences . There are many organism on this earth & humans are also , one of them & it is the immune system which is protecting us , including antibiotics . But the hostile organism can survive them & hit back . It would need a clear vision & pragmatic strategy of humans to develop new drugs to survive in the future .

Well this debate is interesting but probably have some answers in our very basic knowledge, like every bateremia never causes infection, similarly every TB exposure is not giving clinical infections,

1 aspect is the immune mechanisms just containing the organism in a balance

2 aspect is the natural course of every organism with a variety like brucella and lyme disease. and the principle of 33% in those adults who get active TB with out treatment in which 1/3 will clear the infection with out any sequellae, 1/3 will end up in organ damage and 1/3 will die ,, why this difference is just a combination of multiple risk factors as mentioned in the above blogs.

3 incidental granuloma with AFB may not grow on cultures as they may just represent a dead basilli with a immune response.

and the bottom line "low- virulence TB strains" is like a guess for which sky is the limit.

I think Dr. Leonard is right, in that, till an actual study is done to compare and see the difference between strains recovered from patients with active disease versus those that do not have active disease, i think we would all just be guessing and theorizing. One thing is for sure and that is the immune system cannot be separated from the active disease.

In addition as mentioned above, strain virulence may be a factor.

What about a situation where a person normally would contain the infection, but due to continuous exposure to a source, a sort of critical threshold infective dose may overwhelm the immune system. Add to that genetic susceptibility (Africans & African-Americans)....but yes the findings would be very interesting.

From what i understand those who show or do not develop active disease have what is called latent Mtb infection.

This invariably means they are not symptomatic or also known as asymptomatic and are therefore carriers of the infection.

Actually those studies have already been done - and in very large numbers of patients.

We know from - for example - contact tracing studies, that many of the people infected do not develop active disease in the short term. We also know that some of them do. Lastly we know that some of those who are initially infected but healthy develop disease at a later point in life. Here in Copenhagen, we had a patient develop TB 42 years after exposure - and could prove by typing, that it was the same strain his father (his index case) had had. Moreover, it's a strain that has not been circulating for many years in the area, so it is vanishingly unlikely that he was suddenly reinfected with the exact same strain that he had been exposed to 4 decades ago. So here we have an example of the exact same isolate causing both acute and chronic TB.

This is an extreme example in terms of time, but otherwise a perfectly normal one: we know that it is quite usual for a group of people exposed to the same index case, to have some acute cases and a larger number of latent cases, even though the infecting organism is presumably exactly the same.

In addition, there are many studies that have looked at the strains isolated from patient samples and from biopsy material to try and find differences. There are differences between M. tuberculosis strains - our own published work on strain isolation from Madagascar (PMID: 20463103) indicates that "modern" TB strains of the Asian lineage appear to generate slightly different immune responses from the ancient strains and this is associated with slightly higher virulence. It explains why these strains are spreading relatively rapidly. But these strains still produce predominantly latent infection: it's simply a question of degree.

My Ethiopian colleagues have also looked at biopsy material from healthy but infected individuals (PMID: 11191539). The strains identified were later tested by hybridisation and turned out to be those we are familiar with from active disease cases in the area. This result is backed up by the literature: if you perform a pubmed search for "Tuberculosis biopsy PCR" or "Tuberculosis biopsy hybridization" you'll turn up about a thousand articles where the researchers have looked for, or at M. tuberculosis in biopsy material. Many of these studies include typing. And so far, everything indicates that the strains causing latent disease are also those causing active, progressive disease.

Which, when you think about it, makes perfect sense. The inflammation of the lungs in response to M. tuberculosis infection, which causes disease, with cough and infected sputum, is also the method by which the disease spreads. A strain that didn't cause disease, could not spread to new hosts and would go extinct more or less immediately.

The difference between active and latent disease is in the host's immune response. We know that if you treat healthy people who are latently infected with TNF-antagonists, their risks of suddenly developing active TB go way up. A recent case was reported where a lung transplant - containing a latent TB infection - gave rise to acute disease that killed the donor (for a review see PMID: 18926414).

So there's your answer: it's not that there are benign strains of M. tuberculosis. It's that the outcome is rather due to the interaction of the host's immune response and the infecting strain.

I totally agree with Dr. Singh Shivakumar response in regard to latent tuberculosis.

Dr. Garcia-Elorriaga, Hospital of Infectious Diseases, National Medical Center "La Raza", Mexico City.

I think it's worth pointing out also the idea that during latent TB, the disease is entirely inactive is an over-simplification. All "latent TB" actually means is that the disease is not bad enough to drive the patient into care, and that a substantial number of patients with"latent TB" experience TB symptoms of varying degree.

TB is not, in real life, a binary "active or latent" outcome, but a gradient stretching from infection with few, or no symptoms (what we call latent) to acute, fulminant rapidly progressive disease.

When we screened healthy individuals in Ethiopia by X-ray, many of them were TST or quantiferon positive. 30% of them (almost all TST or QFT-positive) had significant scarring and fibrosis in their lungs, consistent with TB (the X-rays were read blinded by a panel of independant specialists) and a substantial minority had features which appeared to be TB-specific features (calcified lesions, etc). Now, you can't have sufficient inflammation in your lungs to generate visible lesions without having symptoms of some kind, but all these patients were quite clear that they had never had TB. The most likely explanation is that they had transient periods of illness which was not identified as tuberculosis.

This is well-documented in the literature - we know from the days of sanatoria that approximately a third of patients who were committed for tuberculosis treatment eventually recovered, without antibiotic treatment. Some of them later relapsed, but the majority thereafter controlled their infection. We also see it in clinical practice - contacts to index cases who are symptomatic, but not yet identified (or self-identified) as TB cases. Some of these will control their disease, and thus be identified subsequently as "latent" disease even though they had active disease at one point (or more than one point). Others will progress.

Some people, of course lie in the middle. They never get full control of the disease and are symptomatic, but don't progress rapidly. Instead they develop a slowly progressive disease that may persist for many, many years, with asymptomatic periods. This too has been well documented in the days of the sanitoria, and we still see it often today in the poorest regions, where TB control problems are weak.

We can also model all of these outcomes in animal models. Non-human primates infected via aerosol with physiological doses can develop progressive and ultimately fatal disease, control the disease with few or no symptoms - or they can fall in between those extremes and develop detectable disease, but with periods of control. Joanne Flynn, who has done the most work in this area, calls these animals "percolators" because their disease "bubbles up" from time to time. They develop pathology that on X-ray looks exactly like our Ethiopian patients.

I think the work from Bouke de Jong is clearly answering your question and the title is self explanatory:

J Infect Dis. 2008 Oct 1;198(7):1037-43. doi: 10.1086/591504.

Progression to active tuberculosis, but not transmission, varies by Mycobacterium tuberculosis lineage in The Gambia.

Thanks to all contributors especially Timothy Doherty. I agree with your statement that latent TB does not mean "entirely inactive". It would be especially relevant in a study looking at biopsy samples from autopsies (of individuals presumed to have died with rather than of TB) and assuming these persons had "benign" TB. It would require robust verification. It may also be difficult to get ethical approval for such a study.

It seems the data being generated from genotyping of myco TB complex strains, if linked to good and reliable clinical data in a bioinformatics model/system, may answer the question to some extent. As previous contributors have suggested, the answers may already be out there.

May I be bold enough to state that any "new" strain of TB which if found to behave in a latent manner in most healthy humans, or cause infection only in severely immunocompromised hosts, be declassified to a status of opportunistic mycobacterium, joining the ranks of MAC etc.

Just a note which covers both the last two comments: the work of the MRC group inthe Gambia confirms that M. africanum is less likely to cause active tuberculosis than M. tuberculosis. It doesn't suggest that different lineages of M. tuberculosis vary significantly in virulence.

So we do know that there are at least 3 species that can commonly cause TB-like disease in humans: M. tuberculosis, M. bovis and M. africanum, in descending order of virulence. But all of these appear to generate a spectrum of disease from latent infection to active disease - it's just that they lie in different places on that spectrum.

Dear Leonard & Timothy, From my perspective it seems reasonable to put progression to active disease under the word virulence and therefore the M. africanum lineage (Lineage 6) would be less virulent than modern TB strains (lineage 4) circulating in the Gambia. However once individuals are infected by one or the other lineage, it needs to reactivate eventually to persist otherwise it would disappear by definition. Now in light of the Gambian study, the transmission rate is the same for both lineages meaning that once an individual is sick independently of the lineage, he will expose his household contacts similarly. However, the proportion of lineage 4 exposed individuals progressing to active disease within the first two years of exposure is significantly higher. Reciprocally, the rate of latent TB infection would logically be lower unless M. africanum is more often entirely cleared by the immune system. So whether the reactivation rate from latent TB infection is also higher in lineage 4 could be be logically expected but not demonstrated yet indeed.

Thanks Damien and Timothy for your answers. We do need to take into account that the situation in The Gambia is a relatively unique one for the country and the others in the region that have M. africanum TB disease. Our recent work (PMCID: PMC3656116, PMCID: PMC3810728 ) provides more justification for Leonard's original post-basically we need to do some more work not ncessarily only on samples from autopsies to gain understanding of M.tb's latency/persistence. Maybe M.africanum related work can show the way since we already see a difference in its natural history, 'lower virulence' that may be related to it slower growth characteristics yet some of its subtypes are sufficiently immunogenic, transmitted and responsible for significant disease.

Thanks indeed to Timothy and others for a sensible and logical exposition of the dose and latency/progression spectrum. Clearly at the population level there are species related characteristics of disease progression/latency and these are affected by immunological factors that are known e.g. HIV/immunosuppression etc. and hyperreactions/dose effects leading to necrotic lesions as well.

However, it seems to me that at the individual patient/contact level we are much less certain of the mechanisms of persistence of low level infection (?perhaps expressed as trivial granuloma) and it is in this are that the original post is important. The contribution here of patient mycobacterial resistance whether genotypically or conceivable environmentally (?Vit D) related is much less well worked out and the original question of whether this is related within mycobacterial species at the strain level as suggested in the original post is still a valid one. To amplify my question earlier, is there any evidence within any RLFP type of M tuberculosis that in family studies (at diagnosis of the index case) any human genotypic factor/profile predicts which family-within-household members will be found to be immunologically reactive without overt features of infection or infected as distinct from non reactive and non infected? Such correlations would relate human genotype to speed of onset of overt infection which in terms of pread must be ragarded as a virulence factor. As regards the original post, a similar question potentially might be asked for organism RFLP type i.e is there any preponderance or dearth of particular human genotypes by the RFLP type - although this may only be answerable in geographically defined areas because of human genotype population diversities and linkages. It will probably only be answerable in low incidence populations where household exposure is the only common mechanism of spread and the population is largely immunocompetent because of the confounding effects of high frequency transmission and multiple sources.

Why are these questions important? Genotypic tests on man and microbe are becoming much easier and more precise targeting either in human populations or with certain strains, and an ability to study the protonomic correlates of either, might enable us to make better progress in human follow-up or strain control whilst we still have effective anti-tuberculous drugs.

I agree that these are important questions, and they've been the subject of intense research for the last 12-13 years.

Right now, there are no clear profiles, but we do have some hints, and the evidence that specific changes in cell wall/cell membrane of M. tuberculosis affect virulence is good (see, for example: http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0023870). In humans, I and some colleagues published several studies identifying specific gene loci in humans that were associated with decreased response to mycobacterial antigen. Unfortunately, followup studies to that were unable to identify a specific linkage to susceptibility or resistance to infection.

The reason is probably that there does not appear to be a clear line between susceptibility and resistance, just as there is not a clear line between latency and active disease. Instead, susceptibility for humans appear to be the outcome of an almost infinite number of factors - age, nutrition, degree of exposure, concurrent infections, genetics, stress, etc. All of these overlap to place an infected person somewhere on the spectrum between eradication and death by TB - and of course, the spectrum is not static. So an infected person can go from latent to active disease and back again (or vice versa) depending on what is happening to them at that point. If you look at the lungs in detail (for example, via PET/HR CT) you can see that a "healthy" person (ie: someone with no symptoms) can actually have multiple active lesions in their lungs where the disease is very much in process. They're just keeping the infection below the threshold where the symptoms become obvious.

For the bacteria, things are a bit more straightforward, and there we have identified many genes directly associated with virulence. But it's important to realise that latency is *also* an important survival strategy for the bacteria. If TB did not establish a latent infection, we would probably already have eliminated it in western Europe and North America. So even genes that don't play a role in direct development of disease play an important evolutionary role for the bacteria, and are thus over the long term, associated with spread and virulence.

"Instead, susceptibility for humans appear to be the outcome of an almost infinite number of factors - age, nutrition, degree of exposure, concurrent infections, genetics, stress, etc. All of these overlap to place an infected person somewhere on the spectrum between eradication and death by TB - and of course, the spectrum is not static. So an infected person can go from latent to active disease and back again (or vice versa) depending on what is happening to them at that point."

I highly agree! I believe this is also one of the fundamental reasons why development of point of care (POC) TB diagnostics has been so challenging. There seems to be a multitude of outcomes of interaction between the host immune system (innate and adaptive) and the sheer number of antigenic markers that the Mtb presents.

We have some cross-sectional data based on a limited number of sera (well-characterized from FIND) that suggest some variables such as history of previous TB and BCG vaccination status that influence antibody production and C-reactive protein level (unpublished, but hopefully will be soon) -- while interesting, I am not sure how to make sense of these findings especially in terms of application for development of POC diagnostics.

I wonder if having data from good animal model studies could tease out a more up-to-date model of infection (and consequently identify more definitive biomarkers of interest) but my slight pessimism would just think it would not extrapolate well to the multitude of variables that a human host would be exposed to. Indeed there have been quite a number of papers that identify good candidates on microarray or mouse models that do not translate well once you test human population samples. Sometimes I wonder if the all mighty "one biomarker to identify them all" even exists for TB... but I suppose we have to keep trying.

I have to agree that I do not think we will find a single biomarker for TB disease. In the case of diseases like HIV, it was relatively simple, because all we were looking for was a marker of infection. Since there's no such thing (as far as we know) as a benign HIV infection, and the disease is still relatively rare, that's good enough to initiate monitoring and care.

With M. tuberculosis, though, around 90% of infected people will not develop disease and the disease is relatively common in endemic regions - in our recent study of over 6000 school children in India, the rate of infection was already over 10% in younger children, and increased with age. So we already have a pretty good marker of infection - but that's not good enough. We simply can't put 20-30% of the whole Indian population on treatment, particularly given the risk of side effects from the drugs used.

A good example of where we are headed, I think is the recognition that TB biomarkers will likely need a 2 or 3 step "risk assessment". So, for example, back in 2001, we published a paper showing that contacts of TB cases who later went on to develop TB themselves almost all had a very high response to ESAT-6. The problem was that not everyone who had a high ESAT-6 response developed disease. Roland Diel published a larger study in TB contacts in 2008 (PMID: 18276940) and figure 2 in that paper illustrates the problem perfectly. Out of 601 TB contacts, 6 later developed disease during 2 years of followup. Of those 6, 100% had a Quantiferon response that was off the top of the scale. So far, so good. The problem is that 22 others also had a Quantiferon response that was off the top of the scale. 6/28 is a lot better than 6/601 when it comes to prediction, but it's still not good enough.

On the other hand, if we can find markers that seperate the 6 high responders who developed disease from the 22 who did not, we can envisage a 2-step testing procedure, whereby a high QFT response puts you in the high risk category, and a second screen then identifies a second risk factor associated with disease. At the TB Biomarkers meeting in Virginia in September this year, two groups presented possible markers to do just that. The markers of disease (for example, elevated CRP) are not TB-specific, but are indicative of active inflammatory processes. But as the second step after a positive Quantiferon, they appeared to offer a very high positive predictive value. I think that's the way the field is going right now: a test for infection followed by a marker of active disease processes.