The 'Nigerian factor' and what it may imply for TB in Africa

In 2012 the first ever national survey of prevalence of infectious TB was conducted in Nigeria...

Nigeria’s the most populous country in the region, and this was the first major African prevalence survey in years.

The results proved to be more than a little shocking: the measured 'crude prevalence of smear-positive TB' was revealed to be 256 cases per 100,000 population. In contrast the last estimated gross prevalence rate (one which included both smear-positive and smear-negative cases) which appeared in the 2013 Global TB Report was 163/100,000. This estimate, however, is not comparable with the survey's result - itwas made with the reckoning that 57% of the estimated prevalent disease was sputum-positive, with the rest being sputum-negative...

So the estimated prevalence of sputum-positive disease for 2012 (as published in the Prevalence Survey) was even lower at 93/100,000 – which is a considerable amount less than 256/100,000.

Which means that it looks like TB has been being underestimated by a factor of nearly three in a country which is both one of the WHO's 22 ‘High Burden Countries’ and also one of the 27 ‘High Burden MDR-TB' ones. As such it should surely have been benefiting from extra attention and support from the WHO and so the implications of this survey should have caused shock waves around the world of medicine, public health and medical science. Very little about it appeared anywhere in the global media on the subject, however, in spite its devastatingly serious epidemiological import.

The survey's results were deemed by the WHO itself to be of “high quality”. As such they warrant being taken seriously. And as such it’s worth considering what they might mean, not just for Nigeria, but also for the rest of the region as well.

One of the most obvious conclusions is that the Case Detection Rate (or CDR – which is the percentage of cases who are being diagnosed in relation to the total estimated epidemic's incidence) was being estimated way too high for the country. Calculations from the survey suggest that as few as 16% of all recent cases can have been being spotted when this was previously being estimated and reported to be 51%. This means that there must be much more infectious disease in the community than was previously thought possible – or which was being used within the calculations which produced all of the other estimates for the country.

Other related estimates have now been recalculated and published in the recent WHO Global Tuberculosis Report 2014 - with incidence rates for Nigeria that are now over 200% higher than in the previous report, prevalence rates that are 100% higher (amounting to a very frightening extra 300,000 current cases in the country), and a terrifying 400% hike in mortality numbers. This last highest percentage rise has been recalculated in consideration of the larger-than-previously-estimated number of untreated cases many of whom must be ending up dead.

But there are implications from this survey, of course, for the wider region as well, particularly given that Nigeria has the highest population of all of the countries in the African region.

It would certainly be rash to assume that Nigeria is typical of the rest of the continent and that these results can simply be extrapolated regionally because in ways it’s not. In some important ways it is, though – the modern African phenomenon of sprawling city slums, for instance, is consistent across almost all of the more populous countries in the region. Worryingly it was found from the survey that the biggest variations in disease (and therefore the biggest discrepancies between survey and estimates) were in these very urban slums. The percentage of HIV co-infection among Nigeria’s TB cases, however, is much lower - it's reckoned to be 25%, whilst it’s 34% across the whole region and over 60% in most of the southern half of the continent. Logically therefore, the real mortality rates can be expected to be a lot more dreadful in southern Africa in similar circumstances. But Nigeria also has areas in its northern regions which are highly politically volatile so, although the WHO's Report doesn’t identify this, it’s probable that health infrastructures in these areas are under more extreme strain than in other parts of the region. It is fair to say that simple conclusions relating to the wider region shouldn't be casually drawn from the Nigerian data.

In actuality the WHO’s Report is largely evasive on exactly what it believes this Survey really does imply though. Without such prevalence surveys there's really only one way of assessing the state of the disease and that's by using estimates. The Report reassures us that such estimates are normally calculated “based mainly on reported case notification data and expert opinion about the levels of under-reporting and under-diagnosis”. What this Survey plainly indicates is that such expert opinion can on occasion be woefully wrong – in this case by a factor of three - but, perhaps understandably, the Global Report chooses not to labour this point or specifically identify it.

We’d like to take a minute to do so however, because very similar evidence that expert opinion concerning the African epidemic can be dramatically wrong exists elsewhere in the recent Global Report. In South Africa the gross estimate of new drug-resistant cases last year was reported to be 6,800; meanwhile the actual notifiednumber of drug-resistant cases in the country (in the same Report) was over 26,000 (or about three times the expert estimate). Frighteningly similarly to the Nigerian survey results, the incidence of South African TB seems to have been being estimated at levels that are three times lower than what they actually are - and in fact it's certainly more than this because this doesn't allow for any missing cases. Like Nigeria, South Africa is also one of those so-called ‘High Burden Countries’ and one of the ‘High Burden MDR-TB Countries that have been warranting particular attention and support from the WHO. South Africa moreover is widely recognised as being a primary furnace not just for the African TB epidemic but for the global one as well.

In the WHO's Report there is actually a reference to the fact that the Nigerian estimates have been wayward though the number used to define the degree of variation seem to have been conservatively selected: the revised best estimate of prevalent Nigerian disease, it says, is now “approximately two times higher than the previous estimate” (choosing therefore not to mention the parallel facts that the incident disease can equally accurately be described as being approximately ‘three times higher’, or mortality ‘four times’). Is this ‘spin’ (which is what we think it looks like)?

One of the revised estimates in the WHO’s new Report also now allows for a 44% estimated increase in TB mortality for the African region specifically because of the Nigerian survey - but at the same time the Nigerian hike allows for just over 400%. This amounts to a nine-fold difference. Is the lower hike for the wider region entirely reasonable given the low rates of HIV co-infection in Nigeria and the number of other large urban slums elsewhere in the region? We think that it may not be.

The regional prevalence rates have also been recalculated, amounting to an increase of less than 4% when the Nigerian shift is 111%; and regional incidence rates have been recalibrated rising by nearly 12% when in Nigeria they have shifted by 328%. Do these huge differences really make coherent sense? We’re far from sure that they do.

The Report does accept that this critical survey has necessitated an upward revision of global estimates of absolute numbers of disease, but at the same time the WHO specifically resists accepting the possibility that this might affect global trends. Is this in any way a correct assessment? We’re not sure – in fact we don’t think it really can be. Africa has the highest rates of TB regionally – so surely such significant revelation must have an impact on global trends? We wonder why on earth the WHO shows itself to be so reluctant to see it this way.