Abstract
Study Design.
Randomized trial.
Objective.
To assess whether infliximab is superior to placebo in reducing Modic change (MC) edema, and whether MC edema or apparent diffusion coefficient (ADC) values of MCs modify the effect of infliximab on disability or low back pain (LBP).
Summary of Background Data.
In the present BackToBasic trial, infliximab did not reduce disability or LBP at five months follow-up in patients with chronic LBP and MC type 1. The effect on MC edema and in edema-defined subgroups is unknown.
Materials and Methods.
Patients with chronic LBP and type 1 MCs were randomized to receive four infliximab infusions or placebo over 98 days. MC edema was assessed using short tau inversion recovery imaging. Primary edema variables were maximum baseline edema volume (Volmax) ≥25% of vertebral body marrow (yes/no) and reduced edema at six months (yes/no). Maximum MC-related ADC value (0%–100%) was measured at baseline. Outcomes at five months were the Oswestry disability index (ODI, 0–100, primary outcome) and LBP intensity (0–10). The analyses included logistic regression and linear mixed-effects models.
Results.
One hundred twenty-eight patients (mean age 43 yr, 84 women) were included, of which 78 were treated per protocol (PP). The odds ratio for reduced MC edema at six months in the infliximab versus placebo group was 2.2 (95% CI: 0.8–5.8; P=0.12) in the primary PP analysis and 2.1 (95% CI: 1.02–4.5; P=0.04) in the total sample. Neither MC edema nor ADC values modified the effect of infliximab on ODI or LBP intensity. At five months, the effect in the Volmax ≥25% group was -4.2 ODI points (95% CI: -11.4 to 3.1; primary PP analysis).
Conclusion.
Infliximab had no clinically relevant edema reducing effect. MC edema did not modify the effect of infliximab on disability or LBP, nor did the MC-related ADC value.
Level of Evidence.
Level II.
Key Words: edema, infliximab, low back pain, lumbar spine, magnetic resonance imaging, Modic changes, prospective studies, randomized controlled trial, STIR, treatment
Modic changes (MCs) are signal changes on magnetic resonance imaging (MRI) extending from the endplate into the vertebral body marrow.1 They are divided into types 1 (MC1, edema), 2 (fatty), and 3 (sclerotic).1,2 MC1 is characterized by edema, inflammatory cells, fibrovascular granulation tissue, and increased nerve fiber density in the endplate.1,3–5 MCs, especially MC1, are thought to be a potential source of chronic low back pain (LBP).6–9 The association between MC1 and LBP was inconsistent in cross-sectional and case-control studies10–13 but was supported by a recent systematic review of longitudinal studies.14
It has been suggested that MCs can be caused by persistent inflammatory stimuli from an autoimmune response.4,15 MC regions contain increased levels of tumor necrosis factor (TNF) and other proinflammatory cytokines.4,16,17 TNF inhibitors like infliximab are used to treat autoimmune and inflammatory diseases and can reduce bone edema.18,19 In ankylosing spondylitis, an effect on edema was evident within six weeks to three months and was also observed at four to seven months and two years.20–24 The pathophysiological mechanism of edema likely differs between ankylosing spondylitis and MCs. One would still expect any relevant effect of infliximab in reducing MC-related edema to appear within a few months.
The recent Norwegian BackToBasic trial investigated the effects of infliximab in patients with chronic LBP and MC1.25 We previously reported that infliximab administered four times over a period of 98 days was no more effective than placebo in reducing disability or LBP at five months follow-up.26 Here, we report the effect on MC-related edema at six months, providing requested additional evidence regarding the role of TNF inhibitors in LBP.27 The effect of infliximab on disability and LBP in edema-defined subgroups is also presented. We assessed MC edema on short tau inversion recovery (STIR) MRI and measured apparent diffusion coefficient (ADC) values of MCs, which are increased by edema.28
The aims of this study were to evaluate (a) whether infliximab is superior to placebo in reducing MC edema at six months; and (b) whether MC edema or ADC values of MCs at baseline modify the effect of infliximab on disability or LBP at five months. We also explored (c) whether these baseline MC features prognosticate disability or LBP at five months independent of treatment.
MATERIALS AND METHODS
Patients were included from five hospital outpatient clinics in Norway in 2018 to 2022, after giving written informed consent.25 All eligibility criteria are detailed in the Supplemental File, Supplemental Digital Content 1, http://links.lww.com/BRS/C701. Required for inclusion were age 18 to 65 years, LBP duration more than six months, mean LBP intensity more than or equal to five on three 0 to 10 numeric rating scales (NRS) and/or Oswestry disability index (ODI) score ≥25, and an MC containing type 1 at any level Th12-S1 confirmed on a baseline study MRI. MCs with height <10% of vertebral body marrow height or diameter ≤5 mm did not qualify for inclusion. Patients with a specific spinal diagnosis (e.g. tumor, fracture, spondylarthritis, infection, spinal stenosis) were excluded. We ruled out active sacroiliitis in all included patients by baseline STIR and T1-weighed MRI.29 The trial, this study, and statistical analysis plans (SAPs) were registered at ClinicalTrials.gov (NCT03704363).
Randomization and Treatment
Patients were randomized in a 1:1 ratio to receive intravenous infusions of biosimilar infliximab (5 mg/kg) or placebo (NaCl) on day 0, 14, 42, and 98. We used a computer randomization procedure stratified by study center and previous participation in the Antibiotics In Modic changes (AIM) study.30–32 Randomization was blocked within each stratum. Block size was not available to investigators. Infliximab and placebo had the same color and look. Patients, investigators, treatment administrators, data analysts, and statisticians were blinded to the treatment allocation.
Patient-Reported Outcome Measures
LBP-related disability was assessed using the validated Norwegian version of the ODI 2.0 (score range: 0–100, higher values indicating worse disability).25,33,34 A between group difference in improvement of 10 ODI points from baseline to five months follow-up was predefined as the minimal clinically important difference.25,35,36 LBP intensity was assessed by the mean score on three 0 to 10 NRSs (current, worst, and average intensity within the last two weeks).25 In this study, we used ODI (primary outcome) and LBP intensity scores reported at baseline, five months (primary follow-up), and nine months (safety follow-up, data used in sensitivity analyses and post hoc analyses, see Analyses section below).
MRI Assessments and Outcome Measures
Standardized baseline 1.5 T lumbar spine MRI was performed at five centers. This study applied sagittal T1-weighted and T2-weighted fast spin echo images (“T1/T2”), sagittal STIR images, and sagittal ADC maps (Supplemental File, Supplemental Digital Content 2–3, http://links.lww.com/BRS/C702, http://links.lww.com/BRS/C703).
Three radiologists (M.D., A.E., and N.V.), with 5 (M.D.) and >10 years of spine MRI experience (A.E. and N.V.), independently evaluated each of the 12 endplate levels Th12 to S1 using all sagittal slices. Blinded to treatment allocation and clinical outcomes, they assessed T1/T2 with STIR and ADC maps closed and evaluated STIR and ADC maps using T1/T2 images for cross-navigation.
Presence, type, and height of MCs were evaluated on T1/T2.37 Intensity and volume of MC edema on STIR (STIR edema) was assessed. The volume was visually categorized and measured in cm3 and as a percentage of the vertebral body marrow volume using manual segmentation. Vertebral body volume was obtained by MD using manual correction of automated segmentation.38 ADC was measured in MCs, normal vertebral body marrow, and cerebrospinal fluid (CSF). The MRI variables and how the conclusive rating was determined in cases of disagreement are detailed in Table 1. Measurements are shown in Figure 1.
TABLE 1.
Description of MRI Variables and Rating Procedures
| T1/T2 variables—rated by MD and AE; conclusive rating: consensus MD/AE or mean value (for height) | |
| MCs | Signal changes in the vertebral bone marrow that extend from the endplate and are not separated from the endplate, not round shaped and abutting the endplate with a smaller base than height (more likely focal fatty marrow or hemangiomas), and not extending through the endplate (Schmorl’s hernias) |
| MC type | Type 1: hypointense on T1- and hyperintense on T2-weighed images. Type 2: hyperintense on T1 and hyperintense or isointense on T2. Type 3: hypointense on T1 and T2. Borderline type 1 versus 2 with near isointense T1 signal is rated as type 2 |
| MC height | Maximum height of the MC (including all MC types) in percentage of vertebral body marrow height measured along the same line as maximum MC height |
| STIR variables—categorical: MD and NV, if they disagreed, conclusive rating by AE; measurements: MD | |
| STIR edema | MC edema on STIR: visibly increased STIR signal compared with normal vertebral body marrow, located in/abutting an MC region on T1/T2 and/or shaped as an MC |
| STIR edema volume | Subjectively rated as <10%, 10%–<25%, 25%–50%, or >50% of the total vertebral body marrow volume, considering the affected area on all images |
| STIRcm3 | Measured STIR edema volume in cm3. Calculated as the summed areas of hand-drawn ROIs containing the whole STIR edema volume at the respective endplate multiplied with (slice thickness+interslice gap) |
| STIRvol% | Measured STIR edema volume in % of vertebral body marrow volume, using manual segmentation |
| STIRint | Mean STIR signal intensity across the whole STIR edema volume. Calculated as (total signal intensity×number of pixels of each edema area summed across all STIR slices)/summed number of pixels of the whole edema volume |
| CSF-STIRint | Mean CSF intensity measured in a circular ROI (diameter 6 mm, 30–40 pixels) in the midsagittal STIR image, or the next image left or right, avoiding non-CSF structures (like nerve roots seen on T1/T2). CSF intensity is measured behind the lower half of the cranial vertebra of the MC affected vertebral unit, if possible (e.g. behind the lower part of L3 in the L3/L4 unit if the MC is superior or inferior to the L3/L4 disc), and otherwise behind the caudal vertebra of the unit or at the next vertebral unit caudally or cranially |
| Body-STIRint | Mean intensity on STIR in normal appearing (on T1/T2) vertebral body marrow, measured in a circular ROI (diameter 7.5 mm, 50–60 pixels) in the central AP third of the vertebral body with the MC, close to the endplate in the opposite part of where the MC is, cranially or caudally, avoiding the central vertebral vein. If the opposite part is not normal, and always when the MC is in S1, the nearest vertebra above is used for measurement, its caudal part is used if possible; otherwise, its cranial part is used. The measurement is first considered in the midsagittal image and then, if necessary, considered in the next image (left or right) before a new location may be considered |
| STIRint% | Calculated as (STIRint–Body-STIRint)×100%/(CSF-STIRint–Body-STIRint) |
| STIRvolint | Volume×intensity of STIR edema calculated as STIRvol%×STIRint%/100% |
| STIR edema change | Change in edema (volume/intensity) subjectively rated as reduced, unchanged, or increased. At each endplate, a change (e.g. “reduced”) is noted if it is present on ≥2 slices and on ≥2 more slices than any opposite change (e.g. “increased”). If there are opposite changes in edema at different endplate levels, a change in the patient is reported if it outweighs any opposite change on ≥2 slices. |
| ADC variables—measurements by MD and NV; mean of their values defined the conclusive value | |
| MC-ADC | Highest mean ADC value measured in a circular ROI (diameter 7 mm, 45–55 pixels) in the MC region, that is, in the vertebral body marrow at the endplate with MC on T1/T2. The ROI is placed in the most hyperintense MC-related region on the ADC map. If the MC region has uniform hyperintensity on the ADC map, the ROI is placed in the MC area with largest height on T1/T2. MC-ADC is measured at all levels showing a hyperintense MC region on the ADC map, but not for MCs with height <10% of vertebral body height or diameter ≤5 mm |
| CSF-ADC | Mean ADC value in the CSF measured in a circular ROI (diameter 7 mm, 45–55 pixels) placed using the same criteria as for CSF-STIR |
| Body-ADC | Mean ADC value in normal appearing (on T1/T2) vertebral body marrow, measured in a circular ROI (diameter 11 mm, 105–115 pixels) placed using the same criteria as for body-STIR |
| MC-ADC% | Calculated as (MC-ADC–body-ADC)×100%/(CSF-ADC–body-ADC) |
Raters MD, AE, and NV are radiologists with 5 (MD) and >10 years spine MRI experience (AE, NV).
ADC indicates apparent diffusion coefficient; CSF, cerebrospinal fluid; MC, Modic change; ROI, region of interest; STIR, short tau inversion recovery.
Figure 1.
Example of STIR edema and ADC measurements. MC1 (arrows) with (A) high signal on T2- and (B) low signal on T1-weighted images. C, Measurements on STIR of mean or total signal intensities (GY) and number of pixels in normal vertebral body marrow, CSF, and MC area with increased STIR signal. D, Measurements on ADC map of mean signal intensities (GY) and pixels in normal vertebral body marrow, CSF, and the region with most intense MC-related ADC signal. Midsagittal slices were used for measurements in CSF and normal vertebral body. ADC indicates apparent diffusion coefficient; CSF, cerebrospinal fluid; GY, gray scale units; MC, Modic change; STIR, short tau inversion recovery.
Primary and secondary baseline variables were predefined as potential treatment effect modifiers and prognostic variables.
Primary variable (categorical):
Volmax: maximum STIR edema volume at any endplate level, visually categorized as ≥25% of vertebral body marrow volume (yes or no).
Secondary variables (continuous):
Volsum: sum of the two largest measured STIR edema volumes at two different endplate levels in percentage of measured vertebral body marrow volume
Volint-sum: sum of two largest measured values for “volume (%)×mean intensity (%)/100%” of STIR edema at two different endplate levels.
ADCmax: maximum MC-related ADC value in percentage (MC-ADC%) at any endplate level (0%=ADC in normal vertebral body marrow, 100%=ADC in CSF).
Volmax was primary variable since such visual edema categorization using edema sensitive MRI series like STIR had documented reliability37,39 whereas measurements of the edema had unknown reliability. ADCmax was added to supplement the edema variables.
A six-month follow-up MRI was obtained at a median of 27 weeks (interquartile range: 25–30 wk) after baseline MRI. The same MRI protocol and scanner were used at both time points in all patients except one. At each endplate level, STIR edema was concluded to be reduced, unchanged, or increased by comparing baseline and follow-up images (Table 1, Figure 2). In cases with opposite conclusive ratings across endplate levels, MD and AE independently rated the patient’s overall edema change, and they agreed in all cases (n=10).
Figure 2.
Example of reduced STIR edema at six months follow-up compared with baseline. A, Reduced MC edema (arrowhead) on STIR at the six months follow-up. B, Baseline edema on STIR was more extensive and intense (arrowhead). MC indicates Modic change; STIR, short tau inversion recovery.
Predefined Hypotheses
Our main research hypotheses were:
Infliximab is superior to placebo for reducing STIR edema from baseline to six months.
A Volmax ≥25% score at baseline predicts less disability (lower ODI score) at five months follow-up in patients receiving infliximab treatment than in those receiving placebo.
The statistical null hypotheses stated no difference. The alternative hypotheses were two-sided.
Analyses
Treatment effects and effect modifiers were analyzed in the per protocol (PP) set (defined in Supplemental File, Supplemental Digital Content 4, http://links.lww.com/BRS/C704). The full analysis set (FAS), including all patients receiving at least one treatment infusion, was used for sensitivity analyses and to assess prognostic ability independent of treatment. All analyses were preplanned in the SAP, except analyses marked post hoc.
The analyses are detailed in Table 2. Effect of infliximab in reducing MC edema was analyzed using logistic regression. Linear mixed effects (LME) models were used to analyze potential modifiers of the treatment effect on ODI and LBP intensity scores at five months and their ability to prognosticate these scores, independent of treatment. Data on ODI and LBP intensity at nine months were used in sensitivity and post hoc analyses (Table 2).
TABLE 2.
Statistical Analyses—Preplanned and Post Hoc
| Analysis, independent variable(s) | Dependent/outcome variable | Analysis set | |
|---|---|---|---|
| Analyses of the effect of infliximab versus placebo in reducing MC edema | |||
| Preplanned—to assess hypothesis 1 | Logistic regression—independent variable: treatment group (infliximab or placebo), adjusted for study center and baseline edema (Volmax) | Reduced MC edema on STIR from baseline to 6 months (yes/no) | PP; FAS for sensitivity analysis |
| Preplanned | As above, adding the interaction term “Volmax×treatment group” to see if baseline edema modifies the treatment effect | As above | Sensitivity analyses in PP and FAS |
| Post hoc | As above—exchanging Volmax with Volsum or Volint-sum as baseline edema variable—in analyses with and without the interaction “baseline edema×treatment group” | As above | As above |
| Analyses of edema reduction versus change in disability and LBP | |||
| Post hoc | Calculated difference (with 95% CI based on t test) in mean improvement of ODI or LBP intensity from baseline to 9 months between patients with and without reduced MC edema on STIR at 6 months | Improvement in ODI or LBP intensity at 9 months | FAS |
| Analyses of potential MRI modifiers of the effect of infliximab versus placebo on disability and LBP | |||
| Preplanned—to assess hypothesis 2 | Linear mixed-effect (LME) model*—independent variables: baseline ODI score, treatment group and Volmax (fixed factors), their interaction, and study center (random factor) | ODI at 5 months | PP; FAS for sensitivity analysis |
| Preplanned | As above—but with baseline LBP intensity score, not ODI | LBP intensity at 5 months | As above |
| Preplanned | As above—with baseline ODI or LBP intensity score—and exchanging Volmax with Volsum, Volint-sum, or ADCmax (one at the time) as potential modifier of treatment effect | ODI or LBP intensity at 5 months | As above |
| Preplanned | As above—with baseline ODI or baseline LBP intensity score—and Volmax, Volsum, Volint-sum, or ADCmax (one at the time) as potential modifier of treatment effect | ODI or LBP intensity at 9 months | Sensitivity analyses in PP and FAS |
| Analyses of the ability of MRI variables to prognosticate disability and LBP, independent of treatment | |||
| Preplanned | LME model*—independent variables: baseline ODI score, treatment group and Volmax (fixed factors), and study center (random factor)† | ODI at 5 months | FAS |
| Preplanned | As above—but with baseline LBP intensity score, not ODI† | LBP intensity at 5 months | FAS |
ADC indicates apparent diffusion coefficient; FAS, full analysis set; LBP, low back pain; MC, Modic change; ODI, Oswestry disability index; PP, per protocol; STIR, short tau inversion recovery.
Volmax, maximum volume of MC edema on STIR ≥25% of vertebral body marrow volume (yes/no). Volsum, sum of two largest measured volumes of MC edema on STIR (continuous variable). Volint-sum, sum of two largest measured values for “volume (%)×mean intensity (%)/100%” of MC edema on STIR (continuous variable). ADCmax, maximum MC-related ADC value (0%–100% scale, continuous variable).
Previous participation in the antibiotics in Modic changes (AIM) study was not adjusted for in any analysis, despite it was a stratification factor in the randomization, as <5% of patients had participated in AIM.
To include the random factor, LME was used, not ANCOVA as stated in the statistical analysis plan.
The interaction between treatment group and Volmax, Volsum, Volint-sum, or ADCmax was insignificant in the analyses of effect modifiers and was therefore omitted from the analyses of prognostic ability.
Reliability of MRI assessments was estimated using Cohen kappa and limits of agreement (LoA). These estimates were based on the total sample, or for measurements on STIR, on 25 consecutive patients (300 endplate levels) independently measured by MD and AE and remeasured by MD after one year. Kappa was interpreted as follows: 0.00 to 0.20 poor; 0.21 to 0.40 fair; 0.41 to 0.60 moderate; 0.61 to 0.80 good; 0.81 to 1.00 very good reliability.40
The significance level was set at 0.05 (two-sided) to minimize type 2 error, as stated in the SAP, both when testing hypotheses 1 and 2 and when performing further explorative analyses. Missing data were not imputed, since only one to two patients lacked data for the primary PP analysis of hypotheses 1 and 2, and further analyses are supplementary.
We used SPSS 29.0 (IBM Corp., Armonk, New York, USA) for logistic regression, STATA 18.0 (StataCorp., College Station, Texas, USA) for LME, and MedCalc 17.6 (MedCalc Software, Ostend, Belgium) for reliability analysis.
RESULTS
A flowchart of the PP and FAS is shown in Figure 3. Table 3 shows baseline characteristics by treatment group. The PP set included 78 patients aged 30 to 64 years (mean: 43.1 yr, 46 women). The FAS included 128 patients aged 26 to 64 years (mean: 43.3 yr, 84 women). The 50 FAS patients excluded from the PP set had discontinued treatment (n=21) or got infusion outside the predefined visit window (±3 d; ±7 d for last infusion) (n=29) (Supplemental file, Supplemental Digital Content 4, http://links.lww.com/BRS/C704). At baseline, the FAS had 270 MCs containing type 1, all with edema on STIR, of which 234 (87%) were located at the four lower endplates (L4–S1).
Figure 3.
Flowchart of study samples and numbers with data. ADC indicates apparent diffusion coefficient; FAS, full analysis set; LBP int, low back pain intensity; ODI, Oswestry disability index; PP, per protocol; STIR, short tau inversion recovery. One of 129 randomized patients were ineligible, leaving 128 patients in the trial, all of whom received at least one treatment infusion and constituted FAS. The numbers with ODI and LBP data are corrected and differ slightly from numbers in the statistical analysis plan.
TABLE 3.
Baseline Characteristics by Treatment Group
| PP set | FAS | |||||||
|---|---|---|---|---|---|---|---|---|
| Infliximab | Placebo | Infliximab | Placebo | |||||
| N=37 | N=41 | N=64 | N=64 | |||||
| Valid N | Value | Valid N | Value | Valid N | Value | Valid N | Value | |
| Age, mean (SD) | 37 | 41.1 (7.9) | 41 | 45.1 (8.9) | 64 | 42.7 (9.1) | 64 | 44.0 (8.9) |
| Sex: women, n (%) | 37 | 21 (57) | 41 | 25 (61) | 64 | 43 (67) | 64 | 41 (64) |
| ODI score (0–100), mean (SD) | 37 | 32.9 (11.5) | 41 | 36.1 (9.8) | 64 | 32.6 (11.8) | 64 | 34.5 (9.6) |
| LBP intensity (0–10), mean (SD) | 37 | 6.3 (1.2) | 41 | 6.6 (1.2) | 64 | 6.3 (1.4) | 64 | 6.5 (1.2) |
| LBP duration months, median (IQR) | 36 | 32 (18.5–60.0) | 41 | 30 (21.0–57.0) | 63 | 36 (18.5–108) | 63 | 36 (24–81.5) |
| Prior disc herniation surgery, yes, n (%) | 37 | 3 (8) | 41 | 5 (12) | 64 | 7 (11) | 64 | 7 (11) |
| Body mass index, mean (SD) | 37 | 25.2 (3.5) | 41 | 25.1 (2.9) | 64 | 24.8 (3.5) | 64 | 25.1 (3.3) |
| Smoking, yes, n (%) | 36 | 1 (3) | 41 | 3 (7) | 63 | 4 (6) | 63 | 4 (6) |
| Physically heavy work/lifting, yes, n (%) | 34 | 11 (32) | 37 | 7 (19) | 58 | 16 (28) | 59 | 13 (22) |
| Levels with MC containing type 1, superior and/or inferior to disc, n (%) | ||||||||
| L1/L2 | 37 | 1 (3) | 41 | 0 | 64 | 1 (2) | 64 | 0 |
| L2/L3 | 37 | 0 | 41 | 4 (10) | 64 | 2 (3) | 64 | 7 (11) |
| L3/L4 | 37 | 4 (11) | 41 | 2 (5) | 64 | 6 (9) | 64 | 4 (6) |
| L4/L5 | 37 | 9 (24) | 41 | 16 (39) | 64 | 19 (30) | 64 | 23 (36) |
| L5/S1 | 37 | 29 (78) | 41 | 27 (66) | 64 | 44 (69) | 64 | 40 (63) |
| Largest MC height (any MC type), measured in % of vertebral body marrow height, n (%) | ||||||||
| <25% | 37 | 0 | 41 | 0 | 64 | 2 (3) | 64 | 0 |
| 25%–50% | 37 | 21 (57) | 41 | 16 (39) | 64 | 32 (50) | 64 | 28 (44) |
| >50% | 37 | 16 (43) | 41 | 25 (61) | 64 | 30 (47) | 64 | 36 (56) |
| Mean (SD) | 37 | 49.1 (13.0) | 41 | 53.9 (12.7) | 64 | 49.8 (14.4) | 64 | 53.8 (13.4) |
| Volmax–maximum visually estimated STIR edema volume in % of vertebral body marrow volume, n (%) | ||||||||
| <25% | 37 | 26 (70) | 41 | 25 (61) | 64 | 43 (67) | 64 | 35 (55) |
| ≥25% | 37 | 11 (30) | 41 | 16 (39) | 64 | 21 (33) | 64 | 29 (45) |
| Volsum–sum of two largest measured STIR edema volumes in %, median (IQR) | 37 | 34.9 (20.9–44.3) | 41 | 34.3 (23.9–55.7) | 64 | 32.9 (18.9–45.4) | 64 | 34.5 (25.4–56.1) |
| Volint-sum–sum of two largest values for “volume (%)×mean intensity (%)/100%” of STIR edema, median (IQR) | 37 | 10.2 (5.6–18.2) | 41 | 11.9 (9.3–22.4) | 64 | 10.3 (6.0–16.7) | 64 | 14.7 (10.0–21.4) |
| ADCmax–maximum MC-related ADC value in %, mean (SD) | 36 | 37.0 (10.6) | 40 | 42.5 (10.9) | 61 | 37.8 (10.8) | 64 | 41.5 (11.1) |
ADC indicates apparent diffusion coefficient; FAS, full analysis set; IQR, interquartile range; LBP, low back pain; STIR edema, MC-related edema on STIR; MC, Modic change; ODI, Oswestry disability index; PP, per protocol; STIR, short tau inversion recovery; Valid N, number of patients with data.
At the primary five months follow-up, ODI and LBP intensity were reported by 98.7% (77/78) of the PP set and 95.3% (122/128) of the FAS (Figure 3). A six-month MRI was obtained in 97.4% (76/78) of the PP set and 96.1% (123/128) of the FAS.
Effect of Infliximab in Reducing MC Edema at Six Months
At 6 months in the PP set, 58.3% of the infliximab group and 40.0% of the placebo group had reduced MC edema on STIR (Table 4). In the primary PP analysis, the odds ratio (OR) for reduced edema in the infliximab group was 2.2 (95% CI: 0.8–5.8; P=0.12). The corresponding OR was 2.1 (95% CI: 1.02–4.5; P=0.04) in the FAS (sensitivity analysis, Table 5).
TABLE 4.
MC Edema on STIR (STIR Edema) at Six Months MRI by Treatment Group
| PP set | FAS | |||||||
|---|---|---|---|---|---|---|---|---|
| Infliximab group, N=37* | Placebo group, N=41* | Infliximab group, N=64* | Placebo group, N=64* | |||||
| Valid N | Value, n (%) | Valid N | Value, n (%) | Valid N | Value, n (%) | Valid N | Value, n (%) | |
| 6mSTIR change—change in STIR edema (intensity and volume) from baseline to 6 months | ||||||||
| Reduced | 36 | 21 (58) | 40 | 16 (40) | 60 | 34 (57) | 63 | 24 (38) |
| Unchanged | 36 | 6 (17) | 40 | 14 (35) | 60 | 13 (22) | 63 | 23 (37) |
| Increased | 36 | 9 (25) | 40 | 10 (25) | 60 | 13 (22) | 63 | 16 (25) |
| 6mSTIRvol change—change in volume of STIR edema from baseline to 6 months | ||||||||
| Reduced | 36 | 19 (53) | 40 | 13 (33) | 60 | 30 (50) | 63 | 21 (33) |
| Unchanged | 36 | 9 (25) | 40 | 16 (40) | 60 | 18 (30) | 63 | 26 (41) |
| Increased | 36 | 8 (22) | 40 | 11 (28) | 60 | 12 (20) | 63 | 16 (25) |
| Largest 6mSTIR edema volume—highest 6 months score (% of vertebral body marrow volume) | ||||||||
| 0† | 36 | 1 (3) | 40 | 1 (3) | 60 | 2 (3) | 63 | 1 (2) |
| 1 (<10%) | 36 | 8 (22) | 40 | 4 (10) | 60 | 20 (33) | 63 | 11 (18) |
| 2 (10%–<25%) | 36 | 20 (56) | 40 | 20 (50) | 60 | 26 (43) | 63 | 26 (41) |
| 3 (25%–50%) | 36 | 7 (19) | 40 | 13 (33) | 60 | 12 (20) | 63 | 21 (33) |
| 4 (>50%) | 36 | 0 | 40 | 2 (5) | 60 | 0 | 63 | 4 (6) |
| Change in score for largest STIR edema volume from baseline to 6 months | ||||||||
| Lower | 36 | 13 (36) | 40 | 5 (13) | 60 | 24 (40) | 63 | 14 (22) |
| Unchanged | 36 | 18 (50) | 40 | 32 (80) | 60 | 29 (48) | 63 | 46 (73) |
| Higher | 36 | 5 (14) | 40 | 3 (8) | 60 | 7 (12) | 63 | 3 (5) |
Six months MRI is lacking in the PP set for 1 of 37 infliximab patients and 1 of 41 placebo patients, and in the FAS for 4 of 64 infliximab patients and 1 of 64 placebo patients.
All patients with zero score had remaining MC, but no MC edema.
FAS indicates full analysis set; MC, Modic change; MRI, magnetic resonance imaging; PP, per protocol; STIR, short tau inversion recovery.
TABLE 5.
Logistic Regression Models for Reduced MC Edema on STIR From Baseline to Six Months
| B (95% CI) | OR (95% CI) | P (Wald test) | AIC | |
|---|---|---|---|---|
| Per protocol (PP) set, n=76 of 78* | ||||
| Crude model | 46.651 | |||
| Treatment, infliximab versus placebo=ref | 0.8 (−0.2, 1.7) | 2.1 (0.8, 5.6) | 0.12 | |
| Model with Volmax | 48.552 | |||
| Treatment | 0.8 (−0.2, 1.8) | 2.2 (0.8, 5.8) | 0.12 | |
| Volmax ≥25%, yes versus no=ref | 0.2 (−0.9, 1.2) | 1.2 (0.4, 3.3) | 0.10 | |
| Model with Volmax and interaction | 50.005 | |||
| Treatment | 0.5 (−0.6, 1.7) | 1.7 (0.5, 5.5) | 0.37 | |
| Volmax ≥25% | −0.2 (−1.5, 0.1) | 0.8 (0.2, 3.3) | 0.81 | |
| Volmax ≥25%×treatment | 0.8 (−1.3, 2.9) | 2.2 (0.3, 18.4) | 0.46 | |
| Full analysis set (FAS), sensitivity analyses, n=123 of 128† | ||||
| Crude model | 66.795 | |||
| Treatment | 0.8 (0.06, 1.5) | 2.2 (1.1, 4.6) | 0.04 | |
| Model with Volmax | 68.435 | |||
| Treatment | 0.8 (0.02, 1.5) | 2.1 (1.02, 4.5) | 0.04 | |
| Volmax ≥25% | −0.2 (−1.0, 0.5) | 0.8 (0.4, 1.7) | 0.55 | |
| Model with Volmax and interaction | 69.739 | |||
| Treatment | 0.5 (−0.4, 1. 5) | 1.7 (0.7, 4.3) | 0.27 | |
| Volmax ≥25% | −0.5 (−1.6, 0.5) | 0.6 (0.2, 1.7) | 0.32 | |
| Volmax ≥25%×treatment | 0.7 (−0.9, 2.2) | 1.9 (0.4, 9.2) | 0.41 | |
Volmax ≥25%, maximum volume of MC edema on STIR at any endplate level visually assessed as ≥25% of vertebral body marrow volume.
All models are adjusted for study center. Results in bold were predefined as main results in the statistical analysis plan.
Six months MRI is lacking in the PP set for 1 of 37 infliximab patients and 1 of 41 placebo patients.
Six months MRI is lacking in the FAS for 4 of 64 infliximab patients and 1 of 64 placebo patients.
AIC indicates Akaike information criterion; MC, Modic change; OR, odds ratio; STIR, short tau inversion recovery.
The edema reducing effect was similar regardless of which baseline edema variable was adjusted for, and baseline edema did not modify the effect (post hoc analyses, Table 5; Supplemental File, Supplemental Digital Content 5–6, http://links.lww.com/BRS/C705, http://links.lww.com/BRS/C706).
Edema Reduction Versus Change in Disability and LBP (Post Hoc Analyses)
Compared with patients with unchanged or increased edema, those with reduced edema at six months improved slightly more in disability and LBP from baseline to nine months, especially in the FAS (Supplemental File, Supplemental Digital Content 7, http://links.lww.com/BRS/C707). At nine months in the FAS, mean improvement differed 7.2 ODI points (95% CI: 3.4–11.9) and 1.0 LBP points (95% CI: 0.1–1.8) between those with and without reduced edema (post hoc unadjusted analysis).
Modifiers of Treatment Effect on Disability and LBP
Baseline MC edema (Volmax ≥25%) did not modify the effect of infliximab versus placebo on ODI at five months in the primary PP analysis (-4.2 points, 95% CI: -11.4 to 3.1; P for interaction 0.33). Results were also negative in the FAS, for effect on LBP intensity at five months, for effect on ODI/LBP at nine months, and for the three secondary potential effect modifiers (Supplemental File, Supplemental Digital Content 8–9, http://links.lww.com/BRS/C708, http://links.lww.com/BRS/C709).
Prognostic Ability
None of the baseline STIR edema variables or ADCmax prognosticated ODI or LBP intensity at the primary five months follow-up, independent of treatment (Supplemental File, Supplemental Digital Content 10, http://links.lww.com/BRS/C710).
Reliability of MRI Assessments
Reliability findings are detailed in Supplemental File, Supplemental Digital Content 11–15, http://links.lww.com/BRS/C711, http://links.lww.com/BRS/C712, http://links.lww.com/BRS/C713, http://links.lww.com/BRS/C714, http://links.lww.com/BRS/C715. Inter-rater reliability (mean kappa) was very good or good for presence of MC (0.95), presence of STIR edema (0.92), presence of MC1 (0.73), visual categorization of STIR edema volume (0.70), and change in STIR edema (0.80). Inter-rater LoA across L4–S1 was for MC-ADC% 2.4±8.1% points. Inter-rater and intrarater LoA across Th12-S1 was for measured STIR edema volume (cm3) 0.8±2.6 and 0.6±1.8, respectively, and for STIR edema intensity (0%–100% scale) 1.6±6.5% points and 1.4±5.1% points, respectively.
DISCUSSION
This is the first trial to evaluate the effect of infliximab on MC-related edema. Infliximab was not superior to placebo for reducing MC edema on STIR at six months follow-up in the primary PP analysis. Neither edema nor ADC values of MCs at baseline modified the effect of infliximab on disability or LBP at five months follow-up, or prognosticated these outcomes, independent of treatment.
In the sensitivity analysis of the larger FAS, infliximab had an unchanged but significant edema reducing effect (OR: 2.1, P=0.04), supporting the involvement of TNF and autoimmunity.4 It is possible that the primary result of no statistically significant edema reducing effect in the PP set was a false negative finding due to small sample size. However, post hoc analyses indicated limited clinical benefits of reduced edema. The 95% CI for the difference in clinical improvement at nine months between those with and those without reduced edema at six months reached 11.9 ODI points and 1.8 LBP points. These figures are close to the border of clinical relevance (10 ODI points, 1.5–2.0 LBP points).36,41
Furthermore, TNF likely had a limited role in mediating LBP despite its potential role in MC edema, as infliximab had no effect on disability or LBP, regardless of the degree of edema. MC edema may be secondary to other pain sources (structural, mechanical) that are not responsive to TNF inhibitors, reflecting the multifactorial mechanisms of chronic LBP.42 Similarly, TNF inhibitors did not improve pain or disability in knee or hand osteoarthritis43,44 but reduced bone marrow lesions in the hand in one study.45 By contrast, in ankylosing spondylitis, inflammation and edema may be a primary source of pain and TNF inhibitors can provide clinical benefit within 2 to 12 weeks.46 Prolonged follow-up would have been unlikely to change the main results of the present study.
The strengths of this study include the randomized design, predefined hypotheses, variables and analyses, standardized MRI, and MRI follow-up rate >96%. Prior data support the reliability and validity of the STIR edema variables used in the primary analysis of hypotheses 1 and 2: edema change47 and visually categorized edema volume (Volmax).32,37 These variables were based on two or three experienced radiologists’ ratings. Compared with prior reports, our inter-rater reliability was better for categorization of edema change on STIR (kappa 0.80 vs. 0.71),47 for categorization of Volmax (kappa 0.70 vs. 0.56),37 and for measurements of MC-ADC% (LoA L4–S1 had upper border 10.5 vs. 15.0)28 (Supplemental File, Supplemental Digital Content 14, http://links.lww.com/BRS/C714).
A limitation of the study is the low statistical power to assess treatment effects in subgroups, which is a common issue in subgroup analyses.48–50 However, in the primary Volmax ≥25% PP subgroup, the effect of infliximab on the primary outcome (ODI) at five months had a 95% CI (-11.4 to 3.1) that made a clinically relevant effect unlikely. Our use of the PP set to assess treatment effects increased the risk of bias, as excluding noncompliant participants can create prognostic differences among the treatment groups.51 In this study we considered the actual effect of adequate infliximab treatment in the PP set to be more relevant than the effect in the FAS. Moreover, results in the FAS were consistent with those in the PP set. Exploration of prognostic ability was limited and did not provide a fully adjusted prognostic model.
A further limitation is that the conclusive values for the secondary edema variables Volsum and Volint-sum were based on only one rater’s measurement of total STIR edema volume. However, the intrarater agreement was acceptable (LoA had an upper border of 2.4 cm3) considering that the increased STIR signal often has diffuse borders.19 No prior reliability data exists for measuring MC edema volume; previously, edema area was measured on one MRI slice.52,53 Finally, we did not repeat the time-consuming manual measurements of MC edema at follow-up, which could have revealed further changes in edema. Automated edema measurements, without any need for manual segmentation, would be valuable for further research, and intensity thresholding is a potential method to achieve this.54
CONCLUSIONS
In patients with chronic LBP and MC1, infliximab had no clinically relevant effect in reducing MC edema at six months follow-up. MC edema and ADC values of MCs at baseline did not modify the effect of infliximab on disability or LBP at five or nine months follow-ups and showed no prognostic ability. These results do not support the use of infliximab as treatment for patients with chronic LBP and MC1 and are in accordance with the clinical results of the trial.
Key Points
The BackToBasic trial showed no effect of infliximab versus placebo on disability or low back pain (LBP) at five months in patients with chronic LBP and Modic changes type 1 (MC1).
It is unknown whether infliximab reduces edema related to Modic changes, or if such edema modifies the effect of infliximab on disability or LBP.
In this sub study of BackToBasic, infliximab had no clinically relevant effect in reducing the edema at six months, assessed using short tau inversion recovery (STIR) imaging.
Baseline edema did not modify the effect of infliximab on disability or LBP at five months.
The results do not support the use of infliximab as a treatment for chronic LBP with MC1.
Supplementary Material
ACKNOWLEDGMENTS
The authors thank their collaborators in BackToBasic study group for their contributions.
Footnotes
This study was approved by the Regional Committee for Medical and Health Research Ethics South East Norway (ref. No. 2017/2450). The device(s)/drug(s) that is/are the subject of this manuscript is/are exempt from FDA or corresponding national regulations because: The Norwegian Medical Products Agency (EudraCT No. 2017-004861-29) approved the use of infliximab in this trial before it started.
This study was funded by the Norwegian National Program for Clinical Therapy Research (KLINBEFORSK), a Governmental Organization (grant No. 2017201).
G.B. has received consulting fees and honoraria as a speaker from Johnson & Johnson. The remaining authors report no conflicts of interest.
Supplemental Digital Content is available for this article. Direct URL citations are provided in the HTML and PDF versions of this article on the journal's website, www.spinejournal.com.
Contributor Information
Magnhild H. Dagestad, Email: magnhild.dagestad@gmail.com.
Nils Vetti, Email: nils.vetti@helse-bergen.no.
Lars C. Haugli Bråten, Email: l.c.h.braten@ous-research.no.
Elisabeth Gjefsen, Email: elgjef@ous-hf.no.
Lars Grøvle, Email: lars.grovle@so-hf.no.
Kristina Gervin, Email: kristina.gervin@farmasi.uio.no.
Anne J. Haugen, Email: anne.julsrud.haugen@so-hf.no.
Gunnstein Bakland, Email: gunnstein.bakland@unn.no.
Gunn H. Marchand, Email: gunn.hege.marchand@helse-midt.no.
Thomas Kadar, Email: thomas.istvan.edelsvard.kadar@helse-bergen.no.
Kjersti Storheim, Email: kjersti.storheim@ous-research.no.
John-Anker Zwart, Email: j.a.h.zwart@medisin.uio.no.
Jörg Assmus, Email: jorg.assmus@helse-bergen.no.
Ansgar Espeland, Email: ansgar.espeland@gmail.com.
References
- 1.Modic MT, Steinberg PM, Ross JS, Masaryk TJ, Carter JR. Degenerative disk disease: assessment of changes in vertebral body marrow with MR imaging. Radiology. 1988;166:193–199. [DOI] [PubMed] [Google Scholar]
- 2.Modic MT, Masaryk TJ, Ross JS, Carter JR. Imaging of degenerative disk disease. Radiology. 1988;168:177–186. [DOI] [PubMed] [Google Scholar]
- 3.Perilli E, Parkinson IH, Truong LH, Chong KC, Fazzalari NL, Osti OL. Modic (endplate) changes in the lumbar spine: bone micro-architecture and remodelling. Eur Spine J. 2015;24:1926–1934. [DOI] [PubMed] [Google Scholar]
- 4.Dudli S, Fields AJ, Samartzis D, Karppinen J, Lotz JC. Pathobiology of Modic changes. Eur Spine J. 2016;25:3723–3734. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ohtori S, Inoue G, Ito T, et al. Tumor necrosis factor-immunoreactive cells and PGP 9.5-immunoreactive nerve fibers in vertebral endplates of patients with discogenic low back Pain and Modic Type 1 or Type 2 changes on MRI. Spine. 2006;31:1026–1031. [DOI] [PubMed] [Google Scholar]
- 6.Albert HB, Kjaer P, Jensen TS, Sorensen JS, Bendix T, Manniche C. Modic changes, possible causes and relation to low back pain. Med Hypotheses. 2008;70:361–368. [DOI] [PubMed] [Google Scholar]
- 7.Maatta JH, Karppinen J, Paananen M, et al. Refined phenotyping of modic changes: imaging biomarkers of prolonged severe low back pain and disability. Medicine. 2016;95:e3495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Splendiani A, Bruno F, Marsecano C, et al. Modic I changes size increase from supine to standing MRI correlates with increase in pain intensity in standing position: uncovering the “biomechanical stress” and “active discopathy” theories in low back pain. Eur Spine J. 2019;28:983–992. [DOI] [PubMed] [Google Scholar]
- 9.Saukkonen J, Maatta J, Oura P, et al. Association between Modic changes and low back pain in middle age: a northern finland birth Cohort study. Spine. 2020;45:1360–1367. [DOI] [PubMed] [Google Scholar]
- 10.Herlin C, Kjaer P, Espeland A, et al. Modic changes-Their associations with low back pain and activity limitation: a systematic literature review and meta-analysis. PLoS One. 2018;13:e0200677. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hopayian K, Raslan E, Soliman S. The association of modic changes and chronic low back pain: a systematic review. J Orthop. 2023;35:99–106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Brinjikji W, Diehn FE, Jarvik JG, et al. MRI findings of disc degeneration are more prevalent in adults with low back pain than in asymptomatic controls: a systematic review and meta-analysis. AJNR Am J Neuroradiol. 2015;36:2394–2399. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Jensen TS, Karppinen J, Sorensen JS, Niinimaki J, Leboeuf-Yde C. Vertebral endplate signal changes (Modic change): a systematic literature review of prevalence and association with non-specific low back pain. Eur Spine J. 2008;17:1407–1422. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Han CS, Maher CG, Steffens D, et al. Some magnetic resonance imaging findings may predict future low back pain and disability: a systematic review. J Physiother. 2023;69:79–92. [DOI] [PubMed] [Google Scholar]
- 15.Dudli S, Liebenberg E, Magnitsky S, Lu B, Lauricella M, Lotz JC. Modic type 1 change is an autoimmune response that requires a proinflammatory milieu provided by the “Modic disc”. Spine J. 2018;18:831–844. [DOI] [PubMed] [Google Scholar]
- 16.Dudli S, Sing DC, Hu SS, et al. ISSLS PRIZE IN BASIC SCIENCE 2017: Intervertebral disc/bone marrow cross-talk with Modic changes. Eur Spine J. 2017;26:1362–1373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Moon HJ, Kim JH, Lee HS, et al. Annulus fibrosus cells interact with neuron-like cells to modulate production of growth factors and cytokines in symptomatic disc degeneration. Spine (Phila Pa 1976). 2012;37:2–9. [DOI] [PubMed] [Google Scholar]
- 18.Marzo-Ortega H, McGonagle D, Rhodes LA, et al. Efficacy of infliximab on MRI-determined bone oedema in psoriatic arthritis. Ann Rheum Dis. 2007;66:778–781. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Bonel HM, Boller C, Saar B, Tanner S, Srivastav S, Villiger PM. Short-term changes in magnetic resonance imaging and disease activity in response to infliximab. Ann Rheum Dis. 2010;69:120–125. [DOI] [PubMed] [Google Scholar]
- 20.Sieper J, Baraliakos X, Listing J, et al. Persistent reduction of spinal inflammation as assessed by magnetic resonance imaging in patients with ankylosing spondylitis after 2 yrs of treatment with the anti-tumour necrosis factor agent infliximab. Rheumatology (Oxford). 2005;44:1525–1530. [DOI] [PubMed] [Google Scholar]
- 21.Wang YX, Griffith JF, Deng M, et al. Vertebral body corner oedema vs gadolinium enhancement as biomarkers of active spinal inflammation in ankylosing spondylitis. Br J Radiol. 2012;85:e702–e708. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Krabbe S, Sorensen IJ, Jensen B, et al. Inflammatory and structural changes in vertebral bodies and posterior elements of the spine in axial spondyloarthritis: construct validity, responsiveness and discriminatory ability of the anatomy-based CANDEN scoring system in a randomised placebo-controlled trial. RMD Open. 2018;4:e000624. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Maksymowych WP, Salonen D, Inman RD, Rahman P, Lambert RG, Group CS . Low-dose infliximab (3 mg/kg) significantly reduces spinal inflammation on magnetic resonance imaging in patients with ankylosing spondylitis: a randomized placebo-controlled study. J Rheumatol. 2010;37:1728–1734. [DOI] [PubMed] [Google Scholar]
- 24.Braun J, Baraliakos X, Golder W, et al. Magnetic resonance imaging examinations of the spine in patients with ankylosing spondylitis, before and after successful therapy with infliximab: evaluation of a new scoring system. Arthritis Rheum. 2003;48:1126–1136. [DOI] [PubMed] [Google Scholar]
- 25.Gjefsen E, Braten LCH, Goll GL, et al. The effect of infliximab in patients with chronic low back pain and Modic changes (the BackToBasic study): study protocol of a randomized, double blind, placebo-controlled, multicenter trial. BMC Musculoskelet Disord. 2020;21:698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Gjefsen E, Braten LC, Ponzi E, et al. Efficacy of a TNF inhibitor in chronic low back pain with Modic changes type 1: the BackToBasic study—a randomized controlled trial. Arthritis Rheumatol. 2024;77:615–623. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Hanberg JS, Ermann J, Katz JN. Modic changes and tumor necrosis factor inhibition: is the door shut? Arthritis Rheumatol. 2025;77:503–505. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Dagestad MH, Vetti N, Kristoffersen PM, et al. Apparent diffusion coefficient values in Modic changes—interobserver reproducibility and relation to Modic type. BMC Musculoskelet Disord. 2022;23:695. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Lambert RG, Bakker PA, van der Heijde D, et al. Defining active sacroiliitis on MRI for classification of axial spondyloarthritis: update by the ASAS MRI working group. Ann Rheum Dis. 2016;75:1958–1963. [DOI] [PubMed] [Google Scholar]
- 30.Storheim K, Espeland A, Grovle L, et al. Antibiotic treatment In patients with chronic low back pain and Modic changes (the AIM study): study protocol for a randomised controlled trial. Trials. 2017;18:596. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Braten LCH, Rolfsen MP, Espeland A, et al. Efficacy of antibiotic treatment in patients with chronic low back pain and Modic changes (the AIM study): double blind, randomised, placebo controlled, multicentre trial. BMJ. 2019;367:l5654. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Kristoffersen PM, Braten LCH, Vetti N, et al. Oedema on STIR modified the effect of amoxicillin as treatment for chronic low back pain with Modic changes-subgroup analysis of a randomized trial. Eur Radiol. 2021;31:4285–4297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Grotle M, Brox JI, Vollestad NK. Cross-cultural adaptation of the Norwegian versions of the Roland-Morris Disability Questionnaire and the Oswestry Disability Index. J Rehabil Med. 2003;35:241–247. [DOI] [PubMed] [Google Scholar]
- 34.Roland M, Fairbank J. The Roland-Morris Disability Questionnaire and the Oswestry Disability Questionnaire. Spine. 2000;25:3115. [DOI] [PubMed] [Google Scholar]
- 35.Ostelo RW, de Vet HC. Clinically important outcomes in low back pain. Best Pract Res Clin Rheumatol. 2005;19:593–607. [DOI] [PubMed] [Google Scholar]
- 36.Ostelo RW, Deyo RA, Stratford P, et al. Interpreting change scores for pain and functional s tatus in low back pain: towards international consensus regarding minimal important change. Spine (Phila Pa 1976). 2008;33:90–94. [DOI] [PubMed] [Google Scholar]
- 37.Kristoffersen PM, Vetti N, Storheim K, et al. Short tau inversion recovery MRI of Modic changes: a reliability study. Acta Radiol Open. 2020;9:2058460120902402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Kaliyugarasan S, Dagestad M, Papalini E, et al. Multi-Center CNN-Based Spine Segmentation from T2W MRI Using Small Amounts of Data. Presented at: 2023 IEEE 20th International Symposium on Biomedical Imaging (ISBI), Cartagena, Colombia; 2023. [Google Scholar]
- 39.Dagestad MH, Toppe MK, Kristoffersen PM, et al. Dixon T2 imaging of vertebral bone edema: reliability and comparison with short tau inversion recovery. Acta Radiol. 2024;65:273–283. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Sim J, Wright CC. The kappa statistic in reliability studies: use, interpretation, and sample size requirements. Phys Ther. 2005;85:257–268. [PubMed] [Google Scholar]
- 41.Suzuki H, Aono S, Inoue S, et al. Clinically significant changes in pain along the Pain Intensity Numerical Rating Scale in patients with chronic low back pain. PLoS One. 2020;15:e0229228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Li W, Gong Y, Liu J, et al. Peripheral and central pathological mechanisms of chronic low back pain: a narrative review. J Pain Res. 2021;14:1483–1494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Persson MSM, Sarmanova A, Doherty M, Zhang W. Conventional and biologic disease-modifying anti-rheumatic drugs for osteoarthritis: a meta-analysis of randomized controlled trials. Rheumatology. 2018;57:1830–1837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Estee MM, Cicuttini FM, Page MJ, Wluka AE, Wang Y. Efficacy of tumor necrosis factor inhibitors in hand osteoarthritis: a systematic review and meta-analysis of randomized controlled trials. Osteoarthr Cartil Open. 2023;5:100404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Kloppenburg M, Ramonda R, Bobacz K, et al. Etanercept in patients with inflammatory hand osteoarthritis (EHOA): a multicentre, randomised, double-blind, placebo-controlled trial. Ann Rheum Dis. 2018;77:1757–1764. [DOI] [PubMed] [Google Scholar]
- 46.van der Heijde D, Dijkmans B, Geusens P, et al. Efficacy and safety of infliximab in patients with ankylosing spondylitis: results of a randomized, placebo-controlled trial (ASSERT). Arthritis Rheum. 2005;52:582–591. [DOI] [PubMed] [Google Scholar]
- 47.Kristoffersen PM, Vetti N, Grovle L, et al. Amoxicillin did not reduce modic change edema in patients with chronic low back pain: subgroup analyses of a randomised trial (the AIM study). Spine. 2023;48:147–154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Sun X, Briel M, Walter SD, Guyatt GH. Is a subgroup effect believable? Updating criteria to evaluate the credibility of subgroup analyses. BMJ. 2010;340:c117. [DOI] [PubMed] [Google Scholar]
- 49.Assmann SF, Pocock SJ, Enos LE, Kasten LE. Subgroup analysis and other (mis)uses of baseline data in clinical trials. Lancet. 2000;355:1064–1069. [DOI] [PubMed] [Google Scholar]
- 50.Cuzick J. Forest plots and the interpretation of subgroups. Lancet. 2005;365:1308. [DOI] [PubMed] [Google Scholar]
- 51.Gupta SK. Intention-to-treat concept: a review. Perspect Clin Res. 2011;2:109–112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Wang Y, Videman T, Niemelainen R, Battie MC. Quantitative measures of modic changes in lumbar spine magnetic resonance imaging: intra- and inter-rater reliability. Spine (Phila Pa 1976). 2011;36:1236–1243. [DOI] [PubMed] [Google Scholar]
- 53.Tibiletti M, Ciavarro C, Bari V, et al. Semi-quantitative evaluation of signal intensity and contrast-enhancement in Modic changes. Eur Radiol Exp. 2017;1:5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Chronaiou I, Thomsen RS, Huuse EM, et al. Quantifying bone marrow inflammatory edema in the spine and sacroiliac joints with thresholding. BMC Musculoskelet Disord. 2017;18:497. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.