Tuberculosis Incidence Trends from 1990 to 2017 Highlight Impact of Drug Resistance: Results from the Global Burden of Disease Study

Objectives This study aimed to determine the global incidence trends of tuberculosis (TB) from 1990 to 2017. Methods Data was obtained from the Global Burden of Disease (GBD) study. The estimated annual percentage changes (EAPCs) were calculated with the age-standardised incidence rate (ASIR) to estimate trends in incidence of TB, including multidrug-resistant TB (MDR-TB) and extensively drug-resistant tuberculosis (XDR-TB). Results Globally, the number of TB cases was 8965.81×10 3 in 2017, with a 9.42% increase since 1990. The ASIR for TB showed a decreasing trend from 1990 to 2017 (EAPC = −1.19, 95% condence interval [CI]: −1.32 to −1.07). Meanwhile, decreasing trends were observed in 162 countries/territories, particularly in Ethiopia and China where EAPCs were −4.51 (95%CI: −5.22 to −3.80) and −4.21 (95%CI: −4.98 to −3.44), respectively. However, obvious increasing trends of MDR-TB cases occurred in areas with low and low-middle sociodemographic indexes (SDI), with EAPCs of 7.97 (95%CI: 2.47 to 13.75) and 6.30 (95%:1.17 to 11.68), respectively. The ASIR for XDR-TB showed pronounced increasing trends globally from 1991 to 2017, with an EAPC of 11.74 (95%CI: 7.50 to 16.16). The largest rising trends of XDR-TB were observed in Kyrgyzstan (EAPC = 31.06, 95%CI: 23.07 to 39.57), followed by Azerbaijan and Uzbekistan. Conclusions There was a decreasing trend for TB incidence worldwide, although it was more pronounced in specic countries and regions. However, the rapidly rising trends of MDR-TB and XDR-TB cases in low and low-middle SDI areas and countries may have an adverse impact on the global control of TB.


Introduction
Tuberculosis (TB) is one of the most important infectious diseases worldwide, and signi cant success in controlling the transmission of the infection has been achieved. However, antituberculosis drug resistance has become an important health concern in recent years (1).
Over the past decades, epidemiological patterns of TB have changed dramatically. In 2008, over 9 million new cases of TB were diagnosed worldwide, and more than one million individuals died (2). In recent years, there has been a declining trend in the global incidence and mortality rates of TB (3). The 2018 Global Tuberculosis Report reported that the incidence of TB fell by about 2% per year during 2000-2017 (4). However, global control of TB is strongly in uenced by antituberculosis drug resistant bacteria (5).
Multidrug-resistant tuberculosis (MDR-TB) is caused by Mycobacterium tuberculosis strains, which are resistant to at least isoniazid (INH) and rifampin (6). In 1994, the World Health Organization (WHO) launched a global program on surveillance of drug-resistant tuberculosis, and the results showed that the occurrence of MDR-TB had increased dramatically worldwide (7). Globally, MDR-TB was estimated to be approximately 460,000 cases in 2017, which accounted for 3.6% of all new cases and 17% of previously treated cases (4). China, India, Russia, and South Africa experienced the highest burden of MDR-TB, accounting for as much as > 60% of all cases globally (8). Furthermore, an estimated 6.7% of MDR-TB patients were also resistant to uoroquinolones and second-line injectable drugs, and these cases are de ned as extensively drug-resistant tuberculosis (XDR-TB) (4,9). Furthermore, this scenario worsened following an inundation of drug use and human immunode ciency virus (HIV) infection, which accelerated the development of drug resistance (10).
Antituberculosis drug resistance has become an increasing threat to global public health. Therefore, it is necessary to track trends in TB burden including MDR-TB and XDR-TB with the latest data from the Global Burden of Disease (GBD) study. In this study, we investigated the incident trends of TB from 1990 to 2017 to facilitate improvement of TB control strategies.

Data source
Data sources for the TB burden were explored using the Global Health Data Exchange (GHDx) query tool (http://ghdx.healthdata.org/gbd-results-tool). We obtained data on the incidence of TB and MDR-TB from 1990 to 2017 strati ed by sex, age, subtype, sociodemographic index (SDI) area, geographic region, and country, whereas data relative to XDR-TB were available from 1991 to 2017. The SDI was classi ed into 5 categories: low, low-middle, middle, high-middle, and high. Data was available for 21 geographic regions and 195 countries/territories worldwide. The methodologies for estimation of disease burden have been described in detail in previous studies (11). Data on the Human Development Index (HDI) were obtained from the United Nations Development Program (http://hdr.undp.org/en/data).

Statistical analysis
The age-standardised rate (ASR) is a necessary and representative index when considering differences in the age structure of multiple populations, and is calculated using the following formula: where a i is the age-speci c rate in the i th age group, w is the number of people (or the weight) in the corresponding i th age group from among the selected reference standard population, and A is the number of age groups.
The estimated annual percentage changes (EAPCs) is a reliable method for describing the magnitude of the trends in ASR (12). A regression line is tted to the natural logarithm of the rates, for example, y = α + βx + ε, where y = ln(ASR) and x = calendar year. The EAPC was estimated as 100×(exp(β) − 1), and its 95% con dence interval was calculated using the linear regression model. Based on the above calculations, trends were assessed as follows: the EAPC value and respective 95%CI > 0 was used to de ne an increasing trend in ASR; and the EAPC value and respective 95% CI < 0 de ned a decreasing trend in ASR; other values indicated that ASR was stable over time. Pearson's correlation analysis was used to detect factors in uencing EAPCs at a national level, including ASR and HDI in 1990 and 2017. Data were analysed using the Statistical Package for Social Sciences (SPSS; version 25.0; SPSS Inc., Chicago, IL, USA). Choropleth maps were drawn using an R program (version 3.6.2).

Results
Trends in incidence of TB Globally, the incidence of TB was 8965.81 × 10 3 (95%UI: 8191.85 × 10 3 to 9820.79 × 10 3 ) in 2017, corresponding to an increase of 9.42% from 1990 ( Table 1). The overall ASIR decreased by an annual average of 1.19% during 1990-2017 (EAPC = − 1.19, 95%CI: −1.25 to − 1.13). The decreasing trend in ASIR of TB was more obvious in females, with the EAPC of − 1.28 (95%CI: −1.35 to − 1.21). When comparing different age groups, the highest increase in number of TB cases was observed in age groups above 70 years, and a decrease in the number of TB cases occurred in age groups under 14 globally (see web-only Supplementary Table S1). Decreasing trends of TB were observed in most SDI areas and geographic regions. All SDI areas showed a decreasing trend of ASIR, particularly in high-middle SDI areas (EAPC = − 2.83, 95%CI: −3.07 to − 2.59). In terms of geographic regions, the largest increase in incidence was found in Central Sub-Saharan Africa (81.66%), followed by Oceania (64.78%) and Eastern Sub-Saharan Africa (62.39%), while the largest decrease was in East Asia (− 49.05%) ( Table 1). The decreasing trends in ASIR of TB occurred in 19 geographic regions, and those with the largest decrease were in East Asia (EAPC = − 4.05, 95%CI: −4.38 to − 3.72), followed by Andean Latin America and highincome North America (Table 1, Figs. 1 and 2). Among the 195 countries/territories, the highest increased change in the incidence of TB cases occurred in the United Arab Emirates (257.45% increase), while the strongest decrease (− 58.62%) was seen in Estonia (see web-only Supplementary Table S2 Figure S3). The EAPCs had a positive association with the HDI in 2017 at the national level (ρ=−0.20, p = 0.01, see web-only Supplementary Figure S5), but not with the ASIR in 1990.
Incidence trends of XDR-TB Globally, the incident cases of XDR-TB increased 5726.19% from 1991, and was 24.47 × 10 3 (95% UI: 17.68-35 × 10 3 ) in 2017 (see web-only Supplementary Table S4). The ASIR of XDR-TB showed an increasing trend (EAPC = 11.74, 95% CI: 9.05 to 14.50) (see web-only Supplementary Table S4, Fig. 1). Increasing changes in the number of incident cases of XDR-TB were found across all age groups globally, especially in individuals aged between 50 and 69 years (Table S1, see web-only Supplementary Figure   S2). Increasing trends in the ASIR for XDR-TB were observed stratifying cases by sex, SDI area, and geographic region from 1991 to 2017, and were particularly evident for the low and low-middle SDI areas, in which EAPCs were 20.30 (95%CI: 16.08 to 24.67) and 13.85 (95%CI: 11.12 to 16.65), respectively (see web-only Supplementary Table S4 and Supplementary Figure S2) Figure S4). No association was found between EAPCs and the ASIR in 1990, and the HDI in 2017 at a national level.

Discussion
In this study, a decreasing trend in the global incidence of TB was observed from 1990 to 2017, likely due to effective strategies of disease prevention and control established in previous decades, such as poverty reduction, improvement in health infrastructure, vaccination programmes, and international cooperation (13). Differences in the incidence of TB existed in countries and regions based on demographic and socioeconomic status. The fastest decrease in ASR for TB occurred in high-middle and high SDI areas. This might also be associated with economic status and health-related behaviour (14). For example, East Asia had the fastest decline in the number of TB cases, which might be explained by a reduction in poverty and under-nutrition combined with increased funding for TB control activities (15). Meanwhile, the largest decrease in trend in ASIR of TB cases was observed in China. The Chinese government had achieved great success in the control of TB due to the revitalization of anti-TB programs in the 1990s. Meanwhile, national initiatives associated with anti-TB strategies were also endorsed, including reducing poverty, improving health infrastructure, and infectious disease management (16). However, an increasing trend still existed in several countries across Eastern Europe and Sub-Saharan Africa, characterised by population expansion, poverty, and poor health infrastructure (17). Several practical strategies, such as extensive scale-up of the availability of health extension workers and the adoption of the Directly Observed Treatment Short-Course strategy (18), probably explain why Ethiopia had the highest decline in ASIR of TB. However, increasing trends in DS-TB also occurred in some countries, such as Norway, a country with a previously low incidence of TB, which was probably attributed to the ageing population (19).
Targets were established for the End TB Strategy, for example, to reduce tuberculosis incidence by 80% in 2030 compared to 2015 (20). However, the rapid development of antituberculosis drug resistance has threatened the targeted strategies for TB control. The numbers of incident cases and ASIR for MDR-TB and XDR-TB have dramatically increased in areas of low and low-middle SDI areas and countries. In these areas and countries, poverty, malnutrition, and overloaded health systems were still the main social factors associated with the incidence of TB (21). The highest increased change in ASIR of MDR-TB and XDR-TB cases occurred in the former Soviet countries of Kyrgyzstan, Azerbaijan, and Uzbekistan, and was associated with TB patients having a higher risk of developing antituberculosis drug resistance (22), and in countries where there had been a lack of proper implementation of the DOTS strategy (23). Furthermore, the situation deteriorated further with the spread of drug use and increased incidence of HIV infection (24,25). In contrast, only seven countries/territories showed a decrease in the trend of ASIR for MDR-TB cases, with the largest decrease being observed in Slovenia, where rigorous TB prevention and treatment programs had been established (26).
The limitations of this study should be considered. The GBD study incidence estimates depended on the quality and quantity of the collected data and on the potential biases from misclassi cation and/or miscoding of disease by different countries, which may have affected the accuracy and robustness of the results. The diagnosis and detection of drug resistance also varied across countries and over time, which may have caused a potential bias. Although age is an important factor, due to the limitations of the ASR formula, this study estimated trends using the percentage change in the incident number across age groups only.

Conclusion
This study found that there was a global decrease in the incidence of TB between 1990 and 2017 across most areas with different SDI and across different geographic regions. However, the rapidly increasing trend in resistance to antituberculosis drugs has had a signi cant impact on the control of TB worldwide.
There remains a huge challenge in the global control of TB, and more e cient prevention and diagnostic strategies and increasing investment to health infrastructure are urgently needed.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analysed during this study are included in this published article and its supplementary information les. Figure 1 Trends in TB burden including TB, MDR-TB, and XDR-TB, globally and strati ed by SDI areas and geographic regions from 1990 to 2017. Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.