The Global Burden of Disease Study is the largest and most comprehensive study to measure the epidemiological levels and trends worldwide [5]. The GBD study used a Bayesian meta-regression tool, DisMod-MR V.2.1, to pool the heterogeneous data. Detailed information about data processing and modeling methods related to LBP is available in Online Supplementary Appendix S1 of the GBD study 2019 [5].

The data sources used to estimate the burden of LBP around the world can be found through the Global Health Data Exchange query tool [17]. The SDI is a composite indicator of development status that is strongly correlated with health outcomes, estimated based on the total fertility rate, educational attainment, and lag distributed income. Human Development Index (HDI) data were obtained from the Human Development Report 2019 [18]. The available HDI data can be matched with incidence among 187 countries and territories.

Overall temporal trends in incidence were assessed by the all-age incidence and age-standardized incidence, as well as the relative percent change of incidence between 1990 and 2019. Incident cases were divided into five age groups: 5–19, 20–44, 45–59, 60–74, and >75 years, and the proportion of incidence in each age group was calculated.

The age-period-cohort (APC) model framework was used to analyze potential trends in incidence by age, period, and birth cohort [19]. The detailed introduction and parameter settings for the APC model can be found in the Methods section of the Supplementary Material.

We also performed correlation analyses between net drifts and the SDI/HDI in 2019. Additionally, a hierarchy cluster analysis was conducted to group the countries and territories into four categories (I: significant decrease, II: minor decrease, III: remained stable or minor increase, IV: significant increase) in terms of their temporal trends in incidence. In addition, we predicted the incidence rates and incident cases of LBP from 2020 to 2029 by sex based on an APC model using the Nordpred R package, taking into account changes in the rate of change and population structure [20]. To facilitate a comparison with predicted outcomes, we calculated the absolute number of new cases that would occur if the rate remained stable (baseline reference), decreased by 1% per year (optimistic reference), and increased by 1% per year (pessimistic reference) based on actual observed rates in 2019. A Bayesian age-period-cohort (BAPC) model using an integrated nested Laplace approximation (INLA) was performed to verify the stability of the predicted results. The INLA and BAPC R packages were used for the sensitivity analyses. All data analyses were performed using R software (version 4.1.3, R Core Team).

Globally, the estimated number of incident cases of MSK disorders was 322.75 million (95% UI 292.67–354.31) in 2019 (Figure 1A). Among all causes of MSK disorders, LBP accounted for >60% of the incidence [223.46 million (95% UI 197.71 to 252.99)], resulting in the highest proportion of disease burden from MSK disorders globally and across all SDI regions.

From 1990 to 2019, the proportion of LBP incident cases relative to all MSK disorders decreased from 73.9% to 69.2% globally. The proportion of LBP incident cases decreased across all SDI regions, with the highest relative decrease in the high-middle SDI region (Figure 1B). LBP resulted in the least pronounced proportional reduction in the low-SDI region.

Over the past 30 years, the number of new LBP cases has increased from 149.29 million (95% UI 131.31–169.17) to 223.46 million (95% UI 197.71–252.99) by 49.67% (95% UI 45.58%–53.34%), as the global population increased from 53.5 billion (95% UI 52.4–54.6) to 77.4 billion (95% UI 74.8–79.9) with an increase of 44.6%. Globally, the all-age incidence rate of LBP was 2887.97 (95% UI 2555.23–3269.71) per 100,000 population in 2019, a growth of 3.49% (95% UI 0.66–6.02), but the age-standardized incidence rate (ASIR) in 2019 was 2748.9 (95% UI 2425.77–3106.89) per 100,000 population, a 13.25% (95% UI -13.87 to −12.64) decline from 1990.

The percentage change in the number of incident cases substantially increased across all SDI regions between 1990 and 2019, with the highest increase [100.62% (95% UI 98.84 to 102.47)] in the low-SDI regions, which has approximately doubled, and the lowest increase in the high-middle-SDI regions [29.75% (95% UI 25.84 to 33.4)]. From 1990 to 2019, the middle-SDI, low-middle-SDI, and low-SDI regions accounted for an increasing proportion of new LBP cases worldwide. Except for the decrease [−6.13% (95% UI −6.96 to −5.26)] in low-SDI regions, the all-age incidence rates increased relatively in all SDI regions, among which the highest was in middle-SDI regions [8.75% (95% UI 4.65 to 12.67)] and the lowest was in high-middle-SDI regions [4.36% (95% UI 1.21 to 7.29)]. Conversely, the percentage change in ASIRs decreased in all SDI regions. The all-age incidence rate is generally lower than the age-standardized rate in SDI regions, except in low SDI and low-middle SDI regions, which indicates that the all-age incidence rate is more suitable to capture the real burden of LBP in low- and middle-income countries.

Globally, a net drift of −0.37% (95% CI −0.40 to −0.34) per year in incidence was derived from the APC model estimates, ranging from −0.48% (95% CI −0.54 to −0.41) per year in low-middle SDI regions to −0.17% (95% CI −0.19 to −0.14) per year in high SDI regions. That is, in the past 30 years, the incidence rates of LBP overall and across all SDI regions have shown a decreasing trend (Supplementary Tables S2, S3; Figure 2; Supplementary Figure S1).

Of 204 countries and territories, 40 had more than 1 million cases of LBP in 2019. China [40.16 million (95% UI 35.37 to 45.81)], India [29.63 million (95% UI 26.01 to 33.74)] and the United States [13.73 million (95% UI 12.30 to 15.29)] were the top three countries, accounting for 37.40% of the global incident cases of LBP. Thirty-six countries showed a moderate increasing trend (net drifts ≥0.0% per year), with Kiribati [0.17% (95% CI −0.26 to 0.61)], Zambia [0.16% (95% CI 0.11 to 0.20)], and Zimbabwe [0.15% (95% CI 0.12 to 0.19)] being the top three. A total of 168 countries and territories showed a decreasing trend in the incidence rate (net drifts <0·0% per year), and the top three countries were India [−0.97% (95% CI −1.24% to −0.70)], Georgia [−0.73% (95% CI −0.80% to −0.66%)], and China [−0.68% (95% CI −0.76% to −0.60%)] (Supplementary Table S2).

Thailand had the highest increase in the all-age incidence rate [52.93% (95% CI 42.76 to 65.12)] with a net drift of 0.099% per year (95% CI 0.082 to 0.115 per year) in incidence rate. In 2019, 25 countries had all-age incidence rates more than 1.5 times higher than the global level, and most of them were countries with high SDI in Southeast, Central, and Eastern Europe. The ASIRs in three countries (Poland, Vanuatu, and Romania) were more than 1.5 times higher than the global level.

While decreasing net drifts across all SDI regions suggest a favorable incidence reduction of LBP, in some low-SDI countries, such as Afghanistan, Angola, and Niger, the population increased substantially (235.21%, 192.09%, and 190.36%, respectively), and the number of incident cases increased significantly (199.89%, 191.27%, and 183.63%, respectively). Similarly, some countries with high SDI, such as Qatar, the United Arab Emirates, and Saudi Arabia, showed dramatic growth in incident cases (676.05%, 605.35%, and 216.33%) and all-age incidence rates (20.59%, 42.88%, and 42.05%), with net drifts in incidence at −0.03% (−0.34 to 0.28), −0.05% (−0.33 to 0.23), and 0.00% (−0.02–0.03), respectively.

Moreover, populous countries such as India and China, which represent typical emerging economies and belong to BRICS, showed a remarkable decrease in net drifts in incidence, ranking in the top three worldwide [India: −0.97% (−1.24 to −0.70); China: −0.68% (−0.76 to −0.60)] with different patterns of percentage change in all-age incidence rates at −6.47% (−9.83 to −3.16) and 0.57% (−5 to −5.63), respectively. We classified the 204 countries and territories into four categories using hierarchical clustering based on changing trends in incidence (Supplementary Figure S2).

These results indicate that the changing trends of LBP incidence were uneven among countries and were not necessarily consistent with expectations based on SDI at the national level. In addition, the changing direction of incidence captured from traditional indicators (such as all-age incidence rate and ASIR) may not be completely consistent with that indicated by net drifts derived from the APC model, which suggests that it is necessary to distinguish period and cohort effects.

Globally, the incidence rates of LBP in all age groups showed decreasing trends ( p < 0.0001), except for no significant change in the >95 years age group [local drifts: 0.03% (95% CI −0.67 to 0.73) per year]. The steepest incidence reduction occurred in the 15–19 years age group [−0.51% (95% CI −0.57 to −0.45) per year], and the decreasing trend weakened with age. For those under 45 years of age, the incidence of LBP in women declined faster than that in men, whereas the opposite was true for those older than 50 years. Similar trends were observed across the different SDI quintiles. Except for the high SDI region, the incidence rates in all regions improved. In the population over 70 years of age, the incidence of women in countries with high SDI has been increasing (Figure 3A). The local drift of incidence for each country or territory is shown in Supplementary Figures S3–S7.

Globally, the incidence of LBP has shifted from the pediatric population (5–19 years) and young population (20–44 years) to the middle-aged population (45–59 years) and elderly population (>60 years) (Figure 3B) (Supplementary Figures S8–S12). This trend was more obvious in regions with middle SDI, high-middle SDI, and high SDI, but not in regions with low SDI. More than 50% of incident cases were concentrated in the young population group. In regions with a higher SDI, the incidence of LBP in the middle-aged and elderly groups was higher. Especially for the >75 years age group, with the increase in SDI, the relative proportion of LBP incidence also increased. These results indirectly reflect the age trends in the LBP population.

The estimates of age, period, and cohort effects by SDI quintile were derived from the APC model. Similar patterns of age effects were found across different SDI regions. The risk of incidence increased with age, peaking in the 80–84 years age group. The risk was the lowest in those aged 5–10 years, suggesting less severe symptoms. Compared with other regions, the high-middle SDI regions showed an overall higher incidence across all age groups. The incidence risk in women was higher than that in men in the overall and SDI quintiles (Figure 4A).

The period effects showed different degrees of incidence risk reduction across the different SDI quintiles. A more favorable relative risk reduction occurred in the low-middle SDI, middle SDI, and high-middle SDI regions. For countries with low SDI and middle SDI, the period effects remained almost constant during 2000–2014, indicating that there was little improvement in the incidence of LBP. Moreover, the period risk for women was lower than that for men during 2005–2019. In each SDI region, the gender difference in period risks between 2015 and 2019 was significantly greater than before. It is worth noting that, although the period risk of high SDI countries was low in the past decade, the relative risk of women showed an unfavorable and significant upward trend, inconsistent with the trends in other SDI regions (Figure 4B).

Globally, the overall risk of successively young-to-middle-aged birth cohorts has decreased. Like the period effects, the decline in cohort effects was more notable in low-middle SDI, middle SDI, and high-middle SDI countries. The incidence of people born after the 1960s in low-middle SDI countries improved significantly, while the relative risk in high SDI countries did not decrease until the 1950 cohort. Compared with individuals born in the 1950 reference cohort, the relative cohort risks of individuals born in the 2010 cohort ranged from 0.88 (95% CI 0.86–0.90) in low SDI countries to 0.74 (95% CI 0.69–0.78) in low-middle SDI countries (Figure 4C). The age, period, and cohort effects of LBP incidence in each country and territory are shown in Supplementary Figures S13–S27.

A negative correlation (R = −0.29, p < 0.001) was observed between the net drifts in incidence and SDI (in 2019) for LBP (Figure 5A). A negative correlation (R = −0.3, p < 0.001) was found between the net drifts in incidence and HDI (in 2019) for LBP (Figure 5B).

We predicted that the ASIR of LBP would continue to decrease slowly at the global level and that the decreasing trends would be stable in the future. The changing trends of the ASIR in females was similar to but significantly higher than that in males (Figure 6A).

Although the ASIRs for LBP have decreased in the past three decades and are expected to continue to decline in the future, the number of incident cases globally will continue to increase in the next decade because of population growth and aging. Globally, the number of new cases would increase from 223 million in 2019 to 253 million in 2029 for both sexes (males: 95–107 million; females: 128–146 million) (Figure 6B). To assess the robustness of our prediction results, we conducted a sensitivity analysis using the BAPC model to predict the future incidence of LBP worldwide. The results of the BAPC model showed consistent trends with the above results, indicating the stability of the prediction results (Figure 6B).

LBP remains a prominent global public health concern with varying temporal trends in incidence geographically. Despite a decrease in ASIRs over the past three decades, the burden persists at a high level. While ASIRs for LBP are anticipated to further decline over the next decade, the aging population will drive a significant increase in incident cases. Raising awareness about LBP and emphasizing disease prevention and control are crucial for mitigating future burdens. There is a pressing need for expanded data collection on individuals with LBP, especially in low-income countries, to facilitate health data refinement and effectively monitor the disease burden.