Quantifying the burden and trends of isoniazid resistant tuberculosis, 1994-2009

Abstract

Background: Quantifying isoniazid resistant (INH-R) tuberculosis (TB) is important because isoniazid resistance reduces the probability of treatment success, may facilitate the spread of multidrug resistance, and may reduce the effectiveness of isoniazid preventive therapy (IPT).

Methodology/principal findings: We used data reported to the World Health Organization between 1994-2009 to estimate the INH-R burden among new and retreatment TB cases. We assessed geographical and temporal variation in INH-R and reported levels in high HIV prevalence countries (≥2%) to understand implications for IPT. 131 settings reported INH-R data since 1994. A single global estimate of the percentage of incident TB cases with INH-R was deemed inappropriate due to particularly high levels in the Eastern European region where 44.9% (95% CI: 34.0%, 55.8%) of incident TB cases had INH-R. In all other regions combined, 13.9% (95% CI: 12.6%, 15.2%) of incident cases had INH-R with the lowest regional levels seen in West/Central Europe and Africa. Where trend data existed, we found examples of rising and falling burdens of INH-R. 40% of high HIV prevalence countries reported national data on INH-R and 7.3% (95% CI: 5.5%, 9.1%) of cases in these settings had INH-R.

Conclusions/significance: Outside the Eastern European region, one in seven incident TB cases has INH-R, while this rises to nearly half within Eastern Europe. Many countries cannot assess trends in INH-R and the scarcity of data from high HIV prevalence areas limits insight into the implications for IPT. Further research is required to understand reasons for the observed time trends and to determine the effects of INH-R for control of TB.

Figure 1. Percentage of incident TB cases with isoniazid resistance.

World map showing the percentage of incident TB cases with isoniazid resistance (INH-R) from the most recent survey in each setting in three time periods: (a) 1994-1999, (b) 2000-2004, (c) 2005-2009. Grey areas indicate those that did not report national data in the time period in question.

Figure 2. Percentage of all, new and retreatment TB cases with isoniazid resistance by WHO region.

Percentage of (a) all TB cases, (b) new TB cases and (c) retreatment TB cases with any resistance to isoniazid, resistance to isoniazid but not rifampin and multi-drug resistance by WHO region (AFR =  African region, AMR =  region for the Americas, EMR =  Eastern Mediterranean region, EUR E =  Eastern European region, EUR WC =  Western and Central European regions, SEAR =  South-East Asian region, WPR =  Western Pacific region). Countrywide data only from the most recent survey available 1994-2009. Blue bars represent any isoniazid resistance, red bars represent isoniazid resistance without rifampin resistance and white bars represent multi-drug resistance.

Figure 3. Linear trends in isoniazid resistance (INH-R) among new TB cases.

We estimated trends in the percentage of new TB cases with INH-R and the estimated number of new TB cases with INH-R per 100,000 population. Settings are grouped by any linear trend (p<0.1) found (“down” or “up”) or “no consistent linear trend” if no linear trend was found. RF = Russian Federation.

Figure 4. Percentage and number of new TB cases with isoniazid-resistant TB in (a) Botswana and (b) Latvia.

Markers indicate reported data. The numbers with INH-R per 100,000 population (pink triangles) are also subdivided into those with MDR (red ‘+’ symbols) and those with INH-R without rifampin resistance (green ‘X’ symbols). Data points for the percentage of TB cases with INH-R are marked by blue squares. Lines indicate fitted linear trends and are only displayed where a significant linear trend was found (p<0.1). The colour of the line matches the colour of the data points to which the linear trend is fitted. Note that scales are different between (a) and (b) and that there are two y-axes, one for the percentage and one for the number of TB cases with INH-R per 100,000 population.

Figure 5. Trends in the number of new TB cases per 100,000 population with INH-R.

Settings were included that provided at least 3 data points from 1994 to 2009. Trends were estimated for two types of INH-R: INH-R with rifampin resistance (multi-drug resistance (MDR)) and INH-R without rifampin resistance. Settings are grouped by any linear trend (p<0.1) found (“down” or “up”) or “no consistent linear trend” if no linear trend was found. RF = Russian Federation.