We can't answer this accurately, since it would be unethical to actually *not* treat today, so there can be no direct comparision. But we do know that in the pre-antibiotic era, the case fatality ratio for TB ranged from 50-60% in adults (higher in young children). You can find plenty of smaller studies indicating a CFR as high as 80-90% in patients with MDR-TB, which at first glance suggests it might be worse than no treatment (though it must be said that the reported CFR in studies of MDR-TB ranges from 5% to 90%, as there are many factors affecting outcome reporting).

But that's a misleading impression, because the patient populations are not comparable.

There's two reasons for that. The first, and biggest joker in the pack is HIV. HIV is highly prevalent in TB cases, and greatly decreases both survival and response to treatment. It's a lethal combination with MDR TB, and of course there's no HIV represented in TB figures from prior to the introduction of antibiotics. The second reason is that patients with MDR-TB are often those patients who failed therapy (for whatever reason) in the early stages. So they are often "selected" from among a population who was already sicker than most TB patients. That means that when we compare CFRs from 70 years ago with those for MDR-TB patients, we are essentially comparing the CFR in the overall population, with a population that is often much sicker when they start therapy.

If we look at data from before anti-retroviral therapy, patients with HIV tended to have 2-3x the CFR of patients without HIV, and this increased to 12-28x in patients with clinical AIDS. So it has a huge effect. In addition, HIV-positive status tends to be associated with social factors (alcohol/drug abuse, homelessness, poverty, etc) that are also associated with poor compliance: so that complicates the numbers even more.

Still, all that said, we can get some estimates from meta-analyses of the data, which reduce some of the study to study variation (for an overview, see http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3124477/). This indicates that the CFR in MDR patients is much higher than in drug-sensitive patients (2-4x higher, depending on where you live) but much lower than the overall CFR in patients not treated with drugs. HIV, of course, makes the CFR worse, but stil doesn't bring it up to pre-antibiotic levels, if good care is available. And of course, in much of the world, it isn't available.

It's still not a very good comparison - medical science today is far better than it was in the era prior to antibiotics, so it would be a real shock if the CFR wasn't better. After all, about half of TB patients used to die during their illness from things other than TB (primarily respiratory infections such as influenza and pneumonia) and we can treat or prevent those today.

Hi..

My thought on this is that fatality ratio may not be same because after 6 month or 9 month therapy if the patient is not cure than we can say that the person may have the resistance to that particular therapy so at that time we can change the protocol accordingly.

While with the untreated TB it is fatal.

Thank you

untreated tuberculosis is less fatal than irregularly treated or MDR TB cases. fatality rate is much higher in XDR and the newly coined TDR-TB cases . un treated fresh diagnosed TB patient has a survival rate of 10 yrs, in comparision MDR and XDR TB the survival is about 3-5 yrs.generally speaking , case fatality rate among the two is difficult to asses and it is influenced by many other factorslike HIV, associated infections, nutritional status, habits and addictions and above all financial status of the person 

one advantage of mdr tb is, it is less transmissible compared to untreated TB 

We can't answer this accurately, since it would be unethical to actually *not* treat today, so there can be no direct comparision. But we do know that in the pre-antibiotic era, the case fatality ratio for TB ranged from 50-60% in adults (higher in young children). You can find plenty of smaller studies indicating a CFR as high as 80-90% in patients with MDR-TB, which at first glance suggests it might be worse than no treatment (though it must be said that the reported CFR in studies of MDR-TB ranges from 5% to 90%, as there are many factors affecting outcome reporting).

But that's a misleading impression, because the patient populations are not comparable.

There's two reasons for that. The first, and biggest joker in the pack is HIV. HIV is highly prevalent in TB cases, and greatly decreases both survival and response to treatment. It's a lethal combination with MDR TB, and of course there's no HIV represented in TB figures from prior to the introduction of antibiotics. The second reason is that patients with MDR-TB are often those patients who failed therapy (for whatever reason) in the early stages. So they are often "selected" from among a population who was already sicker than most TB patients. That means that when we compare CFRs from 70 years ago with those for MDR-TB patients, we are essentially comparing the CFR in the overall population, with a population that is often much sicker when they start therapy.

If we look at data from before anti-retroviral therapy, patients with HIV tended to have 2-3x the CFR of patients without HIV, and this increased to 12-28x in patients with clinical AIDS. So it has a huge effect. In addition, HIV-positive status tends to be associated with social factors (alcohol/drug abuse, homelessness, poverty, etc) that are also associated with poor compliance: so that complicates the numbers even more.

Still, all that said, we can get some estimates from meta-analyses of the data, which reduce some of the study to study variation (for an overview, see http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3124477/). This indicates that the CFR in MDR patients is much higher than in drug-sensitive patients (2-4x higher, depending on where you live) but much lower than the overall CFR in patients not treated with drugs. HIV, of course, makes the CFR worse, but stil doesn't bring it up to pre-antibiotic levels, if good care is available. And of course, in much of the world, it isn't available.

It's still not a very good comparison - medical science today is far better than it was in the era prior to antibiotics, so it would be a real shock if the CFR wasn't better. After all, about half of TB patients used to die during their illness from things other than TB (primarily respiratory infections such as influenza and pneumonia) and we can treat or prevent those today.

I think CFR in active drug resistant TB and untreated TB are two different entities. Therefore it can not be compared. Moreover, rates for drug resistant TB can be calculated using clinical records even though we need to try different treatment regimes to prevent deaths. If we treat them,then CFR becomes much lower than expected. 

On the other hand, we may see the tip of the iceberg of untreated cases unless we do a whole population study. Then if we identify deaths in retrospective studies we may miss whole lot of true cases or there may be misdiagnosis (Some cases of COPD) among deaths, so we will not have a true picture. If we do screening and identify cases it is unethical to not treat and estimate CFR.

As these limitations are different in each situation, we may not compare them. Added to this,  HIV/AIDS may be an issue when calculating CFR of TB as a single disease.

I agree with Timothy.Today would be not ethical, not treat a patient with TB. CFR in MDR-TB is oftenly reported as very high, and may be would be high that undiagnosed and untreated patients in some scenarios, but there is no way today to make such comparisons.

Might augmented virulence of resistant strains be a factor?   MDR/XDR strains that are highly transmitted often have been selected to contain  mutations or polymorphisms that compensate for the fitness costs of drug resistance mutations, and these, or other selected mutations, might increase their fitness or virulence above that of the average drug sensitive strain. This is a difficult question to study, but might be approached by comparing  transmission and clinical course of matched patients with sensitive or resistant strains.

In general, mutations exert a fitness *cost* - so MDR mutants would be expected to be less fit (and therefore less virulent) than drug sensitive strains, which have, after all, been selected for fitness in humans over thousands of years.

The data supports this (http://www.ncbi.nlm.nih.gov/pubmed/22479407) - drug-resistant strains do not compete well with drug sensitive strains in the absence of antibiotic. This is probably why the spread of MDR strains has been slower than was once feared: they are outcompeted by drug sensitive strains in untreated populations.

Of course, that is not necessarily good news, because it simply confirms that there is a substantial population maintaining the high numbers of circulating TB strains, who are not being reached by treatment programmes.

Recent Whole Genome Studies have suggested that the reduced fitness incurred with some antibiotic resistance mutations may be balanced by other secondary mutations that compensate and increase fitness.  It is not clear, however, whether these fitness increasing mutations occur subsequent to the resistance mutations, or existed before the resistance mutations occurred but provided a genomic background that could tolerate the resistance mutations and allow the strain to remain highly virulent and transmissible.

Hi Philip,it is an interesting question, if you think on exploring this think first that it would be unethical not to treat any form of TB

ethics aside drug resistance tb tend to have very high CFR than untreated tb basically because during the time taken for tb to become drug resistance the individual may have been exposed to other opportunity infections which may also have  brought about other pathphysiology. drug resistance TB takes long time to develop hence implies exposures to many health outcomes

I agree with Timothy and it may not be prudent to  compare the fatality rate.

it is different case fatality among untreated may be high than that with ineffective drug  but also we need study with statistics to be accurate

I agree with response of Abdalla Ibrahim. The case fatality rate is different, but it is wise to control for other factors. Because there factors equally important to chemotherapy