Pretreatment Vitamin D Concentrations Do Not Predict Therapeutic Outcome to Anti-TNF Therapies in Biologic-Naïve Patients With Active Luminal Crohn’s Disease

Abstract Background and Aims Vitamin D has a regulatory role in innate and adaptive immune processes. Previous studies have reported that low pretreatment vitamin D concentrations are associated with primary non-response (PNR) and non-remission to anti-TNF therapy. This study aimed to assess whether pretreatment 25-hydroxyvitamin D concentrations predicted PNR and non-remission to infliximab and adalimumab in patients with active luminal Crohn’s disease. Methods 25-Hydroxyvitamin D concentrations were measured in stored baseline samples from 659 infliximab- and 448 adalimumab-treated patients in the Personalised Anti-TNF Therapy in Crohn’s disease (PANTS) study. Cut-offs for vitamin D were deficiency <25 nmol/L, insufficiency 25–50 nmol/L, and adequacy/sufficiency >50 nmol/L. Results About 17.1% (189/1107; 95% CI, 15.0–19.4) and 47.7% (528/1107; 95% CI, 44.8–50.6) of patients had vitamin D deficiency and insufficiency, respectively. 22.2% (246/1107) of patients were receiving vitamin D supplementation. Multivariable analysis confirmed that sampling during non-summer months, South Asian ethnicity, lower serum albumin concentrations, and non-treatment with vitamin D supplementation were independently associated with lower vitamin D concentrations. Pretreatment vitamin D status did not predict response or remission to anti-TNF therapy at week 14 (infliximab Ppnr = .89, adalimumab Ppnr = .18) or non-remission at week 54 (infliximab P = .13, adalimumab P = .58). Vitamin D deficiency was, however, associated with a longer time to immunogenicity in patients treated with infliximab, but not adalimumab. Conclusions Vitamin D deficiency is common in patients with active Crohn’s disease. Unlike previous studies, pretreatment vitamin D concentration did not predict PNR to anti-TNF treatment at week 14 or nonremission at week 54.


Background
By binding to the vitamin D receptor expressed on most immune cells, vitamin D has a key regulatory role in innate and adaptive immune processes. Relevant to the pathogenesis of inflammatory bowel disease (IBD), in animal and in vitro experimental models, vitamin D modulates tight junctions, maintaining intestinal epithelial integrity and regulating hostmicrobiota interactions. 1 Patients with IBD have multiple risk factors for vitamin D deficiency including chronic diarrhea, bile salt malabsorption, dietary restrictions, and reduced sunlight exposure. Consequently, vitamin D deficiency is more common than in the general population, [2][3][4] and while it does not cause IBD, [5][6][7][8] because of the link with active disease, [9][10][11][12] there is considerable interest in the role of vitamin D as an adjunct to IBD therapies. 13 Over the last 3 decades, the anti-TNF monoclonal antibodies, infliximab and adalimumab, have become the most frequently prescribed biologics for immune-mediated inflammatory diseases. Unfortunately, anti-TNF treatment failure in patients with Crohn's disease is common: One-quarter of patients experience primary nonresponse, one-third of responders lose response, and only 40% of patients are in remission at the end of a year. 14,15 Anti-TNF monotherapy, obesity, smoking, disease severity, and the development of antidrug antibodies are associated with low drug concentrations and subsequent anti-TNF treatment failure. 15,16 Carriage of the HLA-DQA1*05 allele confers a 2-fold risk of developing antibodies to anti-TNF treatment. 17,18 In small retrospective studies, vitamin D deficiency has been associated with primary non-response, non-remission, and durability of anti-TNF therapy. [19][20][21] We sought to assess whether pretreatment 25-hydroxyvitamin D concentrations predicted primary non-response and non-remission to infliximab and adalimumab in patients with Crohn's disease.

Study Design
The Personalised Anti-TNF Therapy in Crohn's Disease study (PANTS) is a UK-wide, multicentre, prospective observational cohort reporting the treatment failure rates of the anti-TNF drugs infliximab (originator, Remicade [Merck Sharp & Dohme, UK] and biosimilar, CT-P13 [Celltrion, South Korea]) and adalimumab (Humira [Abbvie, USA]) in anti-TNF-naïve patients with active luminal Crohn's disease. 15 Patients were recruited between February 2013 and June 2016 at the time of first anti-TNF exposure and studied for 12 months or until drug withdrawal. After 12 months, patients were invited to continue follow-up for a further 2 years. Eligible patients were aged ≥6 years with objective evidence of active luminal Crohn's disease involving the colon and/or small intestine. Exclusion criteria included prior exposure to, or contraindications for the use of, anti-TNF therapy. The choice of anti-TNF was at the discretion of the treating physician and prescribed according to the licensed dosing schedule. Study visits were scheduled at the first dose, week 14, and at weeks 30 and 54. Additional visits were planned for infliximab-treated patients at each infusion and for both groups at treatment failure or exit.
For this analysis, we included all patients who had stored serum available from baseline visits and effectiveness outcomes. Patients were excluded from our effectiveness analysis if they had a stoma as Harvey Bradshaw Index (HBI) and short pediatric Crohn's disease activity index (sPCDAI) scores have not been validated in these patient groups. Patients who were recruited into the study with normal prescreening visit 1 fecal calprotectin and C-reactive protein (CRP) levels, and where the only indication for anti-TNF was perianal disease, were also excluded.

Outcomes
Treatment failure endpoints were primary non-response at week 14, non-remission at week 54, and adverse events leading to drug withdrawal.
We used composite endpoints using the HBI in adults and the sPCDAI in children, corticosteroid use, and CRP to define primary non-response ( Figure S1). Remission was defined as CRP of ≤3 mg/L and HBI of ≤4 points (sPCDAI ≤15 in children), without corticosteroid therapy or exit for treatment failure.
Secondary outcomes included anti-TNF drug concentration measured at weeks 14 and 54 and the time to development of anti-TNF antibodies. Drug persistence was defined as the duration of time from initiation of anti-TNF therapy to exit from the study due to treatment failure.
Patients exited the study when they stopped anti-TNF therapy or had an intestinal resection regardless of surgical outcome. They were deemed to be in non-remission for subsequent time points. Patients who declined to participate in the 2-year extension or who exited the study for loss to follow-up, withdrawal of consent, or elective withdrawal of drug, including for pregnancy, were censored at the time of study exit and excluded from the denominator for subsequent analyses.

Clinical Variables and Laboratory Analyses
Variables recorded at baseline by sites were demographics (age, sex, ethnicity, comorbidities, height and weight, and smoking status) and IBD phenotype and its treatments (age at diagnosis, disease duration, Montreal classification, prior medical and drug history, and previous Crohn's diseaserelated surgeries). At every visit, disease activity score, weight, current therapy, and adverse events were recorded.
Blood and stool samples were processed through the central laboratory at the Royal Devon University Healthcare NHS Foundation Trust (https://www.exeterlaboratory.com/) for hemoglobin, white cell count, platelets, serum albumin, CRP, anti-TNF drug and antidrug antibody concentrations, and fecal calprotectin, respectively. Genotyping methods and the genetic analysis have been reported previously. 17

25-Hydoxyvitamin D concentrations
Serum 25-hydoxyvitamin D was measured in baseline samples between January 22, 2020, and March 20, 2020, using the Elecsys 25-hydoxyvitamin electrochemiluminescence immunoassay (Roche) using the Cobas 801 module on the Cobas 8000 analyzer. 22 This competitive electrochemiluminescence assay uses a ruthenium-complexed vitamin D binding protein to capture vitamin D3 (25-OH) and vitamin D2 (25-OH). The local reference range defines vitamin D deficiency as <25 nmol/L, insufficiency as 25-50 nmol/L, and adequacy/ sufficiency >50 nmol/L. Preanalytical stability of serum 25-hydroxyvitamin D following long-term sample storage at up to −40 °C has been demonstrated previously. 23,24 TNF drug-level assays The IDKmonitor free infliximab (K9655) and adalimumab (K9657) drug-level assays permit quantitative measurement of free therapeutic drug in serum. 15 The assays follow a standard ELISA format using a specific monoclonal antidrug antibody fragment as a capture antibody and peroxidase-labeled antihuman IgG antibody as a detection antibody. The measuring range for both assays is 0.8-45 mg/L, with the absence of the drug being defined using a cutoff of <0.8 mg/L.

Drug-tolerant anti-TNF antibody assays
Total antidrug antibody concentrations were measured with IDKmonitor® ELISA assays (Immundiagnostik AG, Bensheim, Germany) performed on the Dynex DS2 ELISA robot (Dynex technologies, Worthing, UK). The Immundiagnostik (IDK) AG (Bensheim, Germany) IDKmonitor infliximab (K9654) and adalimumab (K9651) total antidrug antibody assays allow semi-quantitative measurement of both free and bound antidrug antibodies. 15 A pretreatment acid dissociation step is used to separate antidrug antibodies from the therapeutic antibody. The assay then follows a standard ELISA format using a recombinant therapeutic antibody as a capture and Pretreatment Vitamin D Concentrations and Therapeutic Outcome to Anti-TNF Therapies 3 detection antibody. The positivity thresholds for the infliximab and adalimumab assays are 9 and 6 AU/mL, respectively. 25

Study Size and Statistical Methods
The sample size calculation for the PANTS study has been reported previously. 15 Here we included all patients who had sufficient stored serum from their baseline visit and had outcome data at week 14.
Statistical analyses were undertaken in R 4.1.3 (R Foundation for Statistical Computing, Vienna, Austria). All tests were 2-tailed and P-values of <0.05 were considered significant. We included patients with missing clinical variables in analyses for which they had data and have specified the denominator for each variable. Continuous data are reported as median and interquartile range (IQR), and discrete data as numbers and percentages. We performed univariable analyses using Fisher's exact, Mann-Whitney U, and Spearman's rank tests to identify differences in baseline characteristics between infliximab-and adalimumab-treated patients, and to determine categorical and continuous factors associated with vitamin D levels and the predefined clinical outcomes above. Multivariable logistic regression analyses were used to confirm factors independently associated with vitamin D deficiency. Rates of immunogenicity and drug persistence were estimated using the Kaplan-Meier method, and comparative analyses were performed using univariable and multivariable Cox proportional hazards regression. Further sensitivity analyses using fecal calprotectin at week 54 as an outcome and stratifying the cohort by vitamin D supplementation and/ or corticosteroid treatments at baseline were undertaken.
Similar to the whole cohort, there were significant differences at baseline between the infliximab-and adalimumab-treated patients, including in age, ethnicity, smoking, body mass index, disease duration, and disease behavior. Patients treated with infliximab had more active disease at baseline than patients treated with adalimumab, as evidenced by higher serum CRP and fecal calprotectin concentrations. At the initiation of anti-TNF treatment, immunomodulator use was higher in patients treated with infliximab compared to those treated with adalimumab, but there was no difference in the proportion of patients treated with corticosteroids.  Table 2 and Table S2. Multivariable linear regression analysis confirmed that baseline sampling during non-summer months ( Figure S2), South Asian ethnicity, lower serum albumin concentrations, and nontreatment with vitamin D supplementation were independently associated with lower vitamin D concentrations ( Figure 2).

Baseline Vitamin D Status and Clinical Outcomes
Overall, 17 Pretreatment vitamin D status did not predict response or remission status to anti-TNF therapy at week 14 (primary non-response: infliximab P = .89, adalimumab P = .18; remission: infliximab P = .19, adalimumab P = .38) or nonremission at week 54 (infliximab P = .13, adalimumab P = .58; Figure 3). Overall, there were no differences in median (IQR) vitamin D levels at baseline according to response or remission status at weeks 14 and 54, respectively ( Figure S3).
In patients who continued in the study beyond week 14, there was no difference in drug persistence between patients with vitamin D deficiency (hazard ratio [HR] 0.98 [95% CI,

Sensitivity Analyses
In a subset of 47.1% (520/1107) of patients who had week 54 fecal calprotectin data, we found a weak negative correla- We then excluded from the whole cohort, all patients who were receiving vitamin D supplementation at baseline. Of 773 patients remaining, pretreatment vitamin D status was not associated with primary non-response at week 14 (P = .15) or non-remission at week 54 (P = .26).

Anti-TNF Drug Concentrations and Time to Immunogenicity
We observed a weak positive correlation between pretreatment vitamin D concentration and anti-TNF drug concentrations at week 14 (infliximab: Rho = 0.10, P = .03; adalimumab: Rho = 0.20, P < .001); however, when we included the factors previously associated with week 14 drug level, vitamin D concentrations were not independently associated in our multivariable models ( Figure S4). We did not demonstrate associations with infliximab or adalimumab drug concentrations at week 54.

Key Results
Vitamin D deficiency is common in patients with active Crohn's disease. Unlike previous studies, pretreatment serum 25-hydroxyvitamin D concentrations did not predict primary non-response, non-remission, or anti-TNF drug persistence. Vitamin D deficiency was, however, associated with a longer time to immunogenicity in patients treated with infliximab.

Interpretation
Our observation that 17% and 48% of patients had vitamin deficiency and insufficiency, respectively, is consistent with previous estimates in patients with active IBD and almost double that compared of healthy controls. [3][4][5] Moreover, our model of 25-hydroxyvitamin D concentration confirmed independent associations with baseline sampling during nonsummer months, South Asian ethnicity, lower serum albumin concentrations, and nontreatment with vitamin D supplementation further validates our findings against clinical outcome. Unlike Winter et al., 19 Zator et al., 20 and Xia et al., 21 in their small mixed cohorts, we did not see associations with primary non-response, durability of IBD therapy or remission at week 54, respectively. The major criticisms of these studies are measurement bias (to be included patients needed to have had vitamin D measured); the retrospective assessment of remission; how disease severity was controlled for; and the lack of data relating to concomitant corticosteroid therapy or vitamin D supplementation. 26 This is the first prospective study with a large enough sample size to adjust for potential confounders to examine the association between pretreatment vitamin D status/concentration and clinical outcomes in patients with Crohn's disease treated with anti-TNF therapy. Our negative findings are consistent with the findings from a small (56 Crohn's disease and 12 ulcerative colitis) prospective study reported by Santos-Antunes et al. 27 In our sensitivity analyses, we did not observe any association between pretreatment vitamin D concentrations  and clinical outcomes when we stratified our data by vitamin D supplementation in patients who were treated with corticosteroids at baseline. Our weak association with fecal calprotectin suggests that vitamin D supplementation might have, at best, a modest immunoregulatory role in anti-TNF therapy. Overall, however, and unlike previous reports, our data provide no additional justification for the use of vitamin D supplementation in anti-TNF treatment over current indications. The finding that vitamin D deficiency was lower in infliximab-treated patients is likely to be explained by more active disease at baseline evidenced by a higher serum CRP and fecal calprotectin observed in infliximab-compared to adalimumab-treated patients in this real-world study.
Our observation that the time to immunogenicity was longer in patients with vitamin D deficiency is of interest. Vitamin D has a central role in antigen presentation and T cell function, with effects on immune tolerance in adaptive immune responses. 28 While vitamin D deficiency did not predict primary non-response to anti-TNF treatment, whether low vitamin D levels protect against the development of antidrug antibodies requires further study.

Limitations and Generalizability
We acknowledge the following limitations. We accept that our data would have been strengthened by endoscopic outcomes or cross-sectional imaging. However, in PANTS, 15 we observed a significant association between clinical outcomes at weeks 14 and 54 with fecal calprotectin, which correlates closely with endoscopic findings. In our sensitivity analysis, we did not observe a clinically useful correlation (Rho = −0.09, P = .04) between pretreatment vitamin D concentrations and calprotectin. The addition of cross-sectional imaging would have strengthened our data in those with a disease affecting the small bowel, but less so in those with disease affecting the large bowel only. Because our stored samples are slowly being exhausted, in particular in children, 16 we accept there was some missingness in our cohort. We may have been underpowered to detect associations between vitamin D concentrations and time to the development of anti-adalimumab antibodies, because immunogenicity events were less common and fewer in adalimumab-than in infliximab-treated patients who were vitamin D deficient at baseline.
Our findings are likely to be generalizable to other patients with Crohn's disease, at least at latitudes similar to those in the United Kingdom. It is possible that vitamin D deficiency in patients with IBD at different latitudes will be less prevalent and whether our findings are generalizable in these populations remain unknown. It is perhaps less likely that vitamin D deficiency influences anti-TNF treatment responses in rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis, psoriasis, hidradenitis suppurativa, and uveitis because these conditions are not associated with

Conclusions
Vitamin D deficiency is common in patients with active Crohn's disease. Unlike previous studies, pretreatment serum 25-hydroxyvitamin D concentration did not predict primary