The Incidence, Pathogen Spectrum, and Resistance Profile of Pyogenic Spondylodiscitis

A Nationwide Retrospective Evaluation of G-DRG Data

Background: The incidence of the spinal infection pyogenic spondylodiscitis is rising, particularly in the aging, multimorbid segment of the population. The goal of this study was to analyze nationwide trends in the epidemiology, pathogen spectrum, and antimicrobial resistance profile of this disease in Germany over the period 2005–2023.

Methods: All hospital discharges were analyzed retrospectively using nationwide German Diagnosis Related Groups (G-DRG) data. Hospitalizations with a primary ICD-10-GM diagnosis of spinal osteomyelitis, pyogenic intervertebral disc infection, or discitis, not otherwise specified (M46.2*, M46.3*, M46.4*) were included for analysis. The age-adjusted incidence was calculated using population data from the Federal Statistical Office. Pathogen determinations and resistance phenotypes were ascertained from supplementary pathogen codes.

Results: A total of 151 544 cases were identified. Over the study period, the age-adjusted incidence doubled (from 5.41 to 10.02 cases per 100 000 persons), the patients’ mean age increased from 65.8 to 72.2 years, and the proportion of patients aged 70 years or more rose (from 42.6% to 62.3%). The mean length of hospital stay decreased, while in-hospital mortality doubled (from 3.25% to 6.79%). At the same time, the rate of pathogen determination improved. The proportion of Staphylococcus aureus fell markedly (from 34.5% to 17.2%), while other gram-positive pathogens increased. Gram-negative pathogens remained stable. Resistant S. aureus isolates became rarer, while resistant enterobacterales and enterococci were identified more frequently. There was an overall downward trend in the rate of resistant pathogens in the final 6 years of the study period (from 14.1% in 2017 to 6.9% in 2023).

Conclusion: In Germany, pyogenic spondylodiscitis mainly affects older patients, and the associated in-hospital mortality is rising. Changes in the pathogen spectrum and in resistance patterns indicate the need for corresponding adjustments to empirical treatment regimens.

Cite this as: Amberger S, Hanses F, Lang S, Schmidt NO, Siller S: The incidence, pathogen spectrum, and resistance profile of pyogenic spondylodiscitis: A nationwide retrospective evaluation of G-DRG data. Dtsch Arztebl Int 2026; 123: 433–8.
DOI: 10.3238/arztebl.m2026.0078

Pyogenic spondylodiscitis is a severe infection of the intervertebral disc and adjacent vertebral bodies. Its incidence has risen from 3–6 to 5–12 cases/100 000 inhabitants in several western countries over the past decade (1, 2, 3). Pyogenic spondylodiscitis occurs predominantly in older and comorbid patients, particularly in the presence of diseases such as diabetes, immunosuppression, intravenous drug use, or previous spinal procedures (4). In parallel, demographic aging and increasing multimorbidity are associated with rising mortality (1). Early antimicrobial therapy remains crucial for successful treatment. Historically, Staphylococcus aureus has been the predominant pathogen. Recent data suggest an increasing contribution of other organisms, reflecting healthcare-associated infections and an aging patient population (5, 6, 7). Local and regional variation in pathogen distribution and resistance patterns require empirical regimens that are standardized yet adaptable in order to remain effective (8).

Most evidence comes from single-center or regional series with limited numbers of patients, narrow catchment areas and short follow-up times. This restricts external validity, hampers characterization of long-term resistance trends, and obscures the evolution of gram-negative and implanted device-associated infections at the national level. Nationwide long-term data are therefore needed. The aim of this study was to use nationwide inpatient data to describe the epidemiology and administrative incidence of hospitalized patients with pyogenic spondylodiscitis in Germany from 2005 to 2023 and to analyze changes in the distribution of the causative pathogens, with a specific focus on clinically relevant antimicrobial resistance.

Material and methods

We performed a retrospective analysis of the German Diagnosis Related Groups (G-DRG) hospital discharge database. Hospitalizations with a primary ICD-10-GM diagnosis of spinal osteomyelitis, pyogenic intervertebral disc infection, or unspecified discitis (M46.2*, M46.3*, M46.4*) were included. The microbiological data were analyzed via secondary ICD-10-GM codes for identified pathogens (A49, B95–B98, U80–U83) and supplementary ICD-10-GM U codes denoting clinically relevant antimicrobial resistance. A detailed description of the methods, including statistical analysis, can be found in the eMethods.

Results

Epidemiological trends and outcome data

We identified 151 544 inpatient cases with a primary ICD-10-GM diagnosis of M46.2*, M46.3*, or M46.4* between 2005 and 2023. The annual number of cases increased from 4457 in 2005 to 9967 in 2023, corresponding to a median annual growth rate of 4.68% and a relative increase of 124%. The highest annual count was observed in 2019, with 10 121 cases. The non-adjusted incidence rose from 5.41/100 000 inhabitants in 2005 to 11.94/100 000 in 2023 (an increase of 121%), while the age-adjusted incidence (reference year: 2005) rose to 10.02/100 000 in 2023 (Figure 1, Table).

The mean age increased from 65 years in 2005 to 72 years in 2023 (p < 0.001). The proportion of patients aged ≥ 70 years rose from 42.6% to 62.3% (p < 0.001). The proportion of men increased from 56.4% to 61.5% (p < 0.001), the male-to-female ratio from 1.28:1 to 1.6:1 (Table).

Table

Descriptive characteristics of the study population

The mean length of stay (mLOS) decreased from 25.3 days in 2005 to 22.7 days in 2023 (p < 0.001), a relative decrease of 10.3%. Stratified by age, mLOS fell from 14.5 to 13.4 days in patients aged 0–30 years, from 23.7 to 21.6 days in those aged 30–70 years and from 28.0 to 23.4 days in patients aged ≥ 70 years: (p = 0.003), demonstrating longer hospitalization with increasing age (eMethods, eFigure, Table). Overall, in-hospital mortality across all years and age groups was 5.03%. Mortality rose from 145 in-hospital deaths in 2005 (3.25%) to 677 in 2023 (6.79%) (p < 0.001), representing a relative increase of 109% in non-adjusted in-hospital mortality risk and paralleling the rise in mean age (Figure 2, Table).

Figure 2

Mortality by year associated with the main G-DRG codes for all cases

eFigure

Mean age of hospitalized patients with pyogenic spondylodiscitis by year and mean length of stay (mLOS) for hospitalized patients with pyogenic spondylodiscitis by year, stratified by age group

To explore changes in mortality and potential influencing factors, we analyzed the periods 2005–2009 and 2021–2023 in detail, as comparably fine-detailed additional G-DRG data were available for these years:

Between 2005 and 2009, 809 deaths occurred among 22 153 inpatients, corresponding to in-hospital mortality of 3.65%. The lumbar region was affected most often (57.7%), followed by the thoracic region (20.1%). The non-adjusted in-hospital mortality varied by segment. It was significantly (p < 0.001) highest in the thoracic region (5.1%), followed by the thoracolumbar region (4.4%) and the cervical region (4.1%).

In 2021–2023, 1974 deaths occurred among 27 873 inpatients, corresponding to in-hospital mortality of 7.08%, significantly higher than in 2005–2009 (p < 0.001). The lumbar spine predominated (58.8%), followed by the thoracic spine (18.9%). In this period, non-adjusted in-hospital mortality was highest for multilocular spondylodiscitis (10.3%, p = 0.0028) followed by the thoracolumbar (7.9%) and the lumbar spine (7.2%)

Pathogen spectrum

Pathogens were coded in 95 604 of 151 544 cases (overall documentation rate 58.8%). The rate of pathogens coded increased from 23.6% in 2005 to 89.2% in 2023 (Table).

The relative proportion of S. aureus declined from 34.5% of cases in 2005 to 17.2% in 2023 (mean annual decrease 3.8%), representing the largest absolute reduction among all documented pathogen groups. In contrast, the relative proportion of other or unspecified staphylococci (including coagulase-negative staphylococci [CoNS]) increased from 22.5% to 25.9%, and that of streptococci from 10.1% to 15.4% (mean annual increase 2.8%). The proportion of Escherichia coli and other enterobacterales decreased slightly from 21.4% to 19.6% (Figure 3).

The relative proportion of pseudomonads together with other non-fermenting gram-negative rods decreased slightly from 5.0% to 4.3%. Other gram-positive organisms increased from 1.3% to 4.7% (mean annual increase 8.6%), whereas the group of other specified bacteria dropped slightly (from 3.7% to 2.7%) (Figure 3).

Figure 3

Proportional distribution of pathogen subgroups over time, shown as annual percentages

Overall, S. aureus remained dominant until around 2014, while other or unspecified staphylococci became the most frequently coded pathogen group from 2015 onward. Rather than replacement of gram-positive by gram-negative pathogens, the data suggest an internal shift within the gram-positive spectrum: the proportional decline of S. aureus is offset by incremental gains among other staphylococci subspecies, streptococci, and other gram-positive taxa, whereas classical gram-negative pathogens such as enterobacterales did not expand to the same extent (Figure 3).

Analysis of resistance-coded cases on the basis of U80/U81 (Figure 4) showed the following pattern:

Figure 4

Proportional distribution of resistance-associated pathogens over time, shown as annual percentages

The relative proportion of resistant S. aureus isolates declined from 8.0% in 2005 to 1.7% in 2023 (mean annual decline 6.8%).

In contrast, resistant enterobacterales isolates, including extended-spectrum beta-lactamase (ESBL)-producing and multi-drug resistant gram-negative (MRGN) E. coli, increased from 0.5% to 2.0% (mean annual increase 19.5%), while the relative incidence of resistant enterococci isolates rose from 0.6% to 2.6% (mean annual increase 16.6%). However, both groups showed a noticeable decline during the last 5 years of the study period (Figure 4).

Resistant non-fermenters remained rare, but became less so (increase from 0% to 0.4%). The relative proportion of the pooled group “other resistant bacteria” decreased from 0.7% to 0.3% (Figure 4).

Moreover, rates of resistant pathogens declined over approximately the last 5 years of our study period from 14.1% in 2017 to 6.9% in 2023 (Figure 4).

From 2013 onward, a modification of the G-DRG subcodes permitted additional division of resistant S. aureus isolates into MRSA and non-MRSA isolates. The absolute number of coded MRSA isolates declined from 319 in 2013 to 133 in 2023, while the relative MRSA share within the resistant S. aureus subgroup fell slightly (from 95.5% in 2013 to 88.1% in 2023).

Discussion

Epidemiological trends

This nationwide analysis over a period of 19 years shows that pyogenic spondylodiscitis has become more frequent in the setting of inpatient care with a rising mortality rate, underlining its increasing clinical relevance in recent years. The disease affects an older and increasingly male patient population and now presents with a microbiological and resistance profile that differs from historical assumptions.

The rise in incidence is in line with data from France, the USA, and Japan, where rates have roughly doubled to about 5–12 cases per 100 000 inhabitants (1, 2, 3). This suggests that the increase may reflect structural changes, e.g., aging populations, higher comorbidity burdens, and greater exposure to invasive procedures and implants (1, 7).

Postoperative spondylodiscitis is known to occur in 0.2–3.6% of invasive spinal procedures. The number of such procedures in Germany increased substantially by 82% between 2005 and 2014, so this may represent an important factor in the increasing incidences of spondylodiscitis (9, 10). The male predominance in pyogenic spondylodiscitis may be related partly to sex-specific physiological differences, sex-related differences in susceptibility to systemic infection and sepsis, and differences in the prevalence and severity of relevant comorbidities (11).

The mean age in our cohort reached 72 years in 2023, and the rate of patients aged ≥ 70 years increased markedly; the proportion of men rose from 56.4% to 61.5%. Together with the French, USA, and Japanese cohorts (1, 2, 3), this supports the concept of a “geriatric spinal infection phenotype” in which frailty, vascular disease, diabetes mellitus, renal insufficiency, and prior spinal surgery with placement of implants are frequent and may contribute to diagnostic delay and more severe presentations. Additionally, as discussed below, we observed changes in the pathogen spectrum and a relative increase in multidrug-resistant organisms especially during the first half of the study period, which may have adversely affected outcomes and contributed to higher mortality rates between 2005 and 2021.

In-hospital mortality in our series increased from 3.25% to 6.79%, in line with data from other western countries showing mean mortality of 2.2% to 7.3% (1, 2, 3). The decrease in mLOS from 25.3 to 22.7 days by 2023 may reflect structural changes within the healthcare system, a trend that is also concordant with the results of a pan-European study from 2000 to 2019 (12). Although the observed decrease in mLOS paralleled the increase in mortality rates, the higher mortality is probably multifactorial and more likely to be associated with advanced age, diabetes mellitus, and prior steroid therapy than causally influenced by a shorter mLOS (13). It remains unclear to what extent changes in pathogen spectrum and rate of multidrug-resistant organisms might also have influenced the changes in mortality. However, the trend of slightly decreasing in-hospital mortality rates since 2021 is at least paralleled (and may be partly induced) by the declining overall rates of resistant pathogens during the last 5 years of our study period.

During the study period, the pathogen detection rate increased markedly, from 23.6% in 2005 to 89.2% in 2023. Together with the roughly doubled incidence rates for spondylodiscitis this led to an almost eight times higher absolute number of cases with positive pathogen detection, possibly as a combined result of improved and more stringent microbiological detection techniques and increased accuracy of coding. This also explains the apparent discrepancy between Figure 1 and Figure 3, as relative pathogen proportions presented in Figure 3 have to be interpreted in the context of the markedly increased overall pathogen reporting rates over time. In combination with the refined coding practice for resistant pathogens since 2013, these aspects certainly also led to improvements in the validity of the recording of microbiological resistance patterns.

Figure 1

Annual incidence of spondylodiscitis per 100 000 inhabitants in Germany with ageadjusted incidence rate (reference year: 2005), all age groups combined

Pathogen spectrum

A disease-specific consensus-based national guideline for spondylodiscitis introduced in Germany in 2020 provides structured recommendations for empirical and targeted therapy and defines S. aureus, enterobacterales and streptococci as mandatory empirical targets (14). For most of the 2005–2023 observation period, however, empirical anti-infective regimens were largely driven by local practice. A nationwide survey underscores this heterogeneity, documenting 16 different empirical combination treatments, some of them with insufficient microbiological coverage (15). Across historical cohorts, S. aureus has consistently been the leading cause of pyogenic spondylodiscitis, accounting for 32–67% of culture-positive cases (1, 5, 16, 17, 18, 19, 20). Our nationwide analysis confirms its central role but reveals a marked relative decline from 34.5% in 2005 to 17.2% in 2023, which indicates a structural shift in the pathogen spectrum compared with earlier studies.

In parallel, other staphylococcal isolates (including CoNS) increased from 22.5% to 25.9% and became the most frequently coded group by 2023, in line with published data on CoNS-related infection rates (16, 18, 21).

Streptococcal infections have become increasingly commonly reported in community-acquired and geriatric spondylodiscitis. Contemporary cohorts from Europe and Japan report Streptococcus spp. in 9.5–23% of cases (1, 5, 16, 18, 20, 22), and a recent study from the USA describes a similar trend with a general increase in invasive group A streptococci (23). Our nationwide data also show a relative rise from 10.1% to 15.4%, making streptococci the fastest-rising gram-positive taxon and supporting the concept of a gradual shift from a predominantly S. aureus-driven spectrum to a more heterogeneous gram-positive spectrum. Other gram-positive organisms increased from 1.3% to 4.7%, likely reflecting an aging, heavily comorbid patient population with higher rates of prior spinal implant surgery and improved diagnostics.

Enterobacterales remain the leading gram-negative pathogens. E. coli alone accounts for 4.3–9% of cases in other series and has been linked to inadequate treatment response (19, 21). In our data, the prevalence of enterobacterales remained relatively stable (changing from 21.4% to 19.6%), emphasizing that gram-negative anti-infective coverage remains essential, particularly in high-risk patients.

Pseudomonads and other non-fermenters, reported in 4–12% of cases in recent studies and associated with intravenous drug use and intensive healthcare system exposure (20, 22, 24), remained at around 5% in our cohort, strongly suggesting a substantial healthcare system-associated component of disease burden. Fastidious or rare bacteria remained uncommon (changing from 3.7% to 2.7%), indicating stable low prevalence.

The trends in resistant pathogens show a distinct “polarity”: European surveillance data demonstrate a 31% decrease in MRSA bloodstream infections from 2005 to 2018 (25). This downward trend persisted through 2023 (26), consistent with the observation of a relative decrease in resistant S. aureus from 8.0% to 1.7% in our study. In contrast, resistant enterobacterales increased from 0.5% to 2.0% and resistant enterococci from 0.6% to 2.6%. This trend indicates that the primary danger of resistance in spondylodiscitis is gradually shifting along the pathogen spectrum.

Limitations

This study has limitations inherent to its design: The marked rise in documented pathogen detection most likely reflects improved testing and coding practice rather than a purely biological phenomenon; relative rates in the initial years of the observation period should therefore be interpreted more as conservative estimates, particularly for rare organisms. Furthermore, the dataset contains no information on the specific microbiological diagnostic procedures performed; therefore, we can only speculate that, as mentioned, changes and improvements in diagnostic practice over time contributed to the observed increase in pathogen detection rates (27). However, there were no specific changes in G-DRG reimbursement practice during the study period that would directly explain or greatly bias the rising incidence of spondylodiscitis. De-identification in the source dataset prevents linkage across individual cases and hospital episodes, thereby limiting availability of detailed variables, in-depth multivariable analyses, and dichotomization between conservatively and surgically treated patients. Moreover, due to source data de-identification, it was not possible to reliably capture rare cases of multiple hospitalizations, multiple secondary ICD-10-GM codes, or mixed pathogens. Furthermore, U80/U81 codes were interpreted as indicators of clinically relevant resistance only during the indexed hospital stay. No care settings outside the DRG system (outpatient, rehabilitation, and day-care facilities) could be captured.

Conflict of interest statement
FH has received financial support from the German Federal Ministry for Education and Research, Network of University Medicine. He has been paid for giving lectures and training courses by the German Society of Emergency Medicine (DGINA), the Academy of Infectiology (Akademie für Infektionsmedizin), the Bavarian Medical Association, and Heraeus.

The remaining authors declare that no conflict of interests exists.

Manuscript submitted on 31 December 2025, revised version accepted on 29 April 2026

Corresponding author
Prof. Dr. med. Sebastian Siller

References

1. Conan Y, Laurent E, Belin Y, et al.: Large increase of vertebral osteomyelitis in France: A 2010–2019 cross-sectional study. Epidemiol Infect 2021; 149: e227

2. Akiyama T, Chikuda H, Yasunaga H, Horiguchi H, Fushimi K, Saita K: Incidence and risk factors for mortality of vertebral osteomyelitis: A retrospective analysis using the Japanese diagnosis procedure combination database. BMJ Open 2013; 3: e002412

3. Issa K, Diebo BG, Faloon M, et al.: The epidemiology of vertebral osteomyelitis in the United States from 1998 to 2013. Clin Spine Surg 2018; 31: E102–8

4. Dayer R, Marco G de, Vazquez O, et al.: Laboratory diagnostics for primary spinal infections in pediatric and adult populations: A narrative review. N Am Spine Soc J 2023; 16100270

5. Mylona E, Samarkos M, Kakalou E, Fanourgiakis P, Skoutelis A: Pyogenic vertebral osteomyelitis: A systematic review of clinical characteristics. Semin Arthritis Rheum 2009; 39: 10–7

6. Lopez J, Tatar Z, Tournadre A, et al.: Characteristics of spontaneous coagulase-negative staphylococcal spondylodiscitis: A retrospective comparative study versus staphylococcus aureus spondylodiscitis. BMC Infect Dis 2017; 17: 683

7. Kehrer M, Pedersen C, Jensen TG, Lassen AT: Increasing incidence of pyogenic spondylodiscitis: A 14-year population-based study. J Infect 2014; 68: 313–20

8. Pantel T, Mende KC, Stangenberg M, et al.: Regional spondylodiscitis disparities: Impact on pathogen spectrum and patients. J Clin Med 2024; 13: 2557

9. Gerometta A, Bittan F, Rodriguez Olaverri JC. Postoperative spondilodiscitis. Int Orthop 2012; 36: 433–8

10. Petzold T, Tesch F, Adler JB, Günster C, Niethard FU, Schmitt J: 10-Jahres-Entwicklung operativer Eingriffe an der Wirbelsäule in Deutschland [10-year development of spine surgery in Germany—an analysis of health insurance funds data from 2005 to 2014]. Z Orthop Unfall 2018

11. Schoenfeld AJ, Reamer EN, Wynkoop EI, Choi H, Bono CM: Does patient sex affect the rate of mortality and complications after spine surgery? A Systematic Review. Clin Orthop Relat Res 2015; 473: 2479–86

12. Golinelli D, Sanmarchi F, Toscano F, Bucci A, Nante N: Analyzing the 20-year declining trend of hospital length-of-stay in European countries with different healthcare systems and reimbursement models. Int J Health Econ Manag 2024; 24: 375–92

13. Joerger AK, Albrecht C, Lange N, Meyer B, Wostrack M: In-hospital mortality from spondylodiscitis: Insights from a single-center retrospective study. J Clin Med 2023; 12: 7228

14. Deutsche Gesellschaft für Orthopädie und Unfallchirurgie: Diagnostik und Therapie der Spondylodiszitis. S2k-Leitlinie. Berlin: Arbeitsgemeinschaft der Wissenschaftlichen Medizinischen Fachgesellschaften (AWMF). https://register.awmf.org/de/leitlinien/detail/151–001 (last accessed on 10 November 2025).

15. Lang S, Walter N, Neumann C, et al.: Aktuelle Praxis der empirischen Antibiotikatherapie bei Spondylodiszitis [Current practice of empiric antibiotic treatment for spondylodiscitis]. Orthopadie (Heidelb) 2022; 51: 540–6

16. Ziarko TP, Walter N, Schindler M, Alt V, Rupp M, Lang S: Risk factors for the in-hospital mortality in pyogenic vertebral osteomyelitis: A cross-sectional study on 9753 patients. J Clin Med 2023; 12

17. Conti J, Geremia N, Di Bella S, et al:: Infectious spondylodiscitis: Epidemiology, diagnosis, microbiological findings, clinical features and outcomes in a 14-year retrospective study. Eur J Clin Microbiol Infect Dis 2025; 44: 3067–78

18. McHenry MC, Easley KA, Locker GA: Vertebral osteomyelitis: Long-term outcome for 253 patients from 7 Cleveland-area hospitals. Clin Infect Dis 2002; 34:1342–50

19. Geisler Crone C, Mose Tetens M, Bengaard Andersen A, Obel N, Lebech AM: Clinical characteristics of pyogenic vertebral osteomyelitis, and factors associated with inadequate treatment response. Int J Infect Dis 2021; 108487–93

20. Kim DY, Kim UJ, Yu Y, et al.: Microbial etiology of pyogenic vertebral osteomyelitis according to patient characteristics. Open Forum Infect Dis 2020; 7: ofaa176

21. Hijazi MM, Siepmann T, El-Battrawy I, et al.: The importance of the bacterial spectrum in the clinical diagnostics and management of patients with spontaneous pyogenic spondylodiscitis and isolated spinal epidural empyema: A 20-year cohort study at a single spine center. BMC Infect Dis 2024; 24: 39

22. Zuluaga-García JP, Leon-Aldana S, Herrera D: Spondylodiscitis: A comprehensive review of diagnostic challenges, microbial etiology, and management strategies. SN Compr Clin Med 2025; 7: 1–12

23. Dunne EM, Hutton S, Peterson E, et al.: Increasing incidence of invasive group a streptococcus disease, Idaho, USA, 2008–2019. Emerg Infect Dis 2022; 28: 1785–95

24. Tsantes AG, Papadopoulos DV, Vrioni G, et al.: Spinal infections: An update. Microorganisms 2020; 8: 476

25. Gagliotti C, Högberg LD, Billström H, et al.: Staphylococcus aureus bloodstream infections: Diverging trends of meticillin-resistant and meticillin-susceptible isolates, EU/EEA, 2005 to 2018. Euro Surveill 2021; 26: 2002094

26. WHO Regional Office for Europe and European Centre for Disease Prevention and Control: Surveillance of antimicrobial resistance in Europe, 2022 data: executive summary. Copenhagen: WHO Regional Office for Europe; 2023.

27. Stangenberg M, Mende KC, Mohme M, et al.: Influence of microbiological diagnosis on the clinical course of spondylodiscitis. Infection 2021; 49: 1017–27