Low BASDAI score alone is not a good predictor of anti-tumor necrosis factor treatment efficacy in ankylosing spondylitis: a retrospective cohort study

Ankylosing spondylitis (AS) is characterised by inflammatory back pain which can lead to spinal structural damage. In addition, AS may also involve the peripheral joints, eyes, skin, and the intestinal systems. Non-steroidal anti-inflammatory drugs are typically the primary treatment option for AS, although anti-tumor necrosis factor (anti-TNF) agents have emerged as a novel alternative treatment [1]. Unlike other diseases, the severity of AS disease cannot be assessed using simple or direct “gold standard” measures, as is the case for glycosylated haemoglobin in diabetes mellitus and blood pressure in hypertension, due to the range and complexity of AS symptoms [2]. Although, several attempts have been made to develop a novel biomarker of AS disease activity and some interesting biomarkers have been identified [3, 4], composite scoring systems, namely the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) and/or Ankylosing Spondylitis Disease Activity Score (ASDAS) have been mainly applied when assessing AS disease activity [5, 6].

The BASDAI contains six questions addressing fatigue, back pain, peripheral joint pain/swelling, enthesitis, and morning stiffness [5]. Similar to the BASDAI, the ASDAS includes self-reported indices of back pain, duration of morning stiffness, peripheral joint pain/swelling, and patient global assessment of disease activity; however, the ASDAS also includes laboratory test results, such as the C-reactive protein (CRP) or erythrocyte sedimentation rate (ESR), and each parameter is weighted, not simply added up as for the BASDAI [6]. Although ASDAS might be affected infections or conditions influencing inflammatory markers, the ASDAS has been considered more objective than the BASDAI and was recently recommended as the “target” scale for AS treatment [7, 8]. Nevertheless, the BASDAI is still widely used because it is simpler and less time-consuming than the ASDAS. Moreover, the BASDAI is used to measure disease activity to initiate or maintain TNF inhibitor therapy in routine clinical practice and is a standard parameter when the cost of anti-TNF therapy for Korean AS patients is covered under the Korean National Health Insurance Service [9].

In some AS patients with low disease activity as measured by the BASDAI, the patient may nonetheless complain of pain or have elevated inflammatory markers. In these cases, assessing disease activity with other measurement tools such as the ASDAS can be helpful. In some such cases, we may find that some patients have both high ASDAS and low BASDAI scores at the same time.

Inaccurate measurement of disease activity can lead to inappropriate treatment and/or medication of AS patients. Therefore, the aim of this study was to determine the prevalence of high disease activity as measured using the ASDAS-CRP, in AS patients who nonetheless have low BASDAI scores after anti-TNF treatment. The clinical impact of this discordance in disease activity on anti-TNF survival was also investigated.

In this retrospective cohort study, we found that 38.8% of patients with low BASDAI scores (< 4) after 3 months (6 weeks for infliximab) of anti-TNF treatment had high disease activity according to ASDAS-CRP assessment (≥ 2.1). These patients tended to have more severe disease activity after 9 months of anti-TNF treatment and greater risk of discontinuation of anti-TNF treatment than those with low disease activity according to both ASDAS-CRP and BASDAI scores. Female sex was also associated with greater ASDAS-CRP values after 3 months (6 weeks for infliximab) of anti-TNF treatment.

There have been few previous studies address in AS disease activity as measured by different assessment systems; however, some studies investigating AS patients treated with anti-TNF agents have provided valuable data. Consistent with our results, one previous study indicated that 36.9% (210/568 patients) of patients with low BASDAI scores (< 4) had high disease activity as measured by the ASDAS-CRP (≥ 2.1) [11]. Moreover, even higher percentages of patients having high ASDAS-CRP (≥ 2.1) among those having low BASDA score (< 4) were shown in two other studies; 62.1% (64/103 patients) and 65.8% (48/73 patients) [12, 13].

Assessing more patients into worse conditions cannot be simply judged as better measurement. Interestingly, we found differences in prognosis between patient groups based on the ASDAS-CRP assessment. Patients with high ASDAS-CRP appeared to have higher disease activity at their follow-up visits despite anti-TNF treatment. Similar results were obtained in previous studies in which the immediate response to the first anti-TNF agent was shown to largely determine the long-term prognosis [14]. In addition, the anti-TNF agent retention rate was significantly lower in AS patients with high ASDAS-CRP results than among patients with low ASDAS-CRP results, although all would be considered to have low disease activity based on their BASDAI score.

“Treat-to-target (T2T)” is a major principle in treating many diseases, such as hypertension, diabetes mellitus, and dyslipidemia. In the past decade, promotion of T2T has spread to inflammatory rheumatic diseases, including rheumatoid arthritis, spondyloarthritis (SpA), systemic lupus erythematosus, and gout [15‐18]. After determining the desired target values, the objectives of T2T are to ensure accurate and reliable disease assessment and more reasonable treatment decisions (i.e., whether treat or not treat and switch or maintain medications), and to thereby obtain better outcomes. In view of the potential adverse effects and financial burden of anti-TNF treatment, a T2T approach should be emphasised in AS patients treated with anti-TNF agents.

Several attempts have been made to compare the discriminatory capacity of the ASDAS with that of BASDAI in AS patients treated anti-TNF agents. One previous study investigating eligibility criteria for initiation of anti-TNF treatment in AS patients showed that an ASDAS-CRP result ≥ 2.1 is associated with a greater likelihood of improvement than a high BASDAI score (≥ 4), but the patients who had both ASDAS-CRP ≥ 2.1 and BASDAI ≥ 4 showed the greatest improvement [12]. Similarly, from another prospective cohort study, applying the ASDAS definition of high disease activity leads to more axial SpA patients being selected to start anti-TNF treatment and these ‘captured’ patients have a greater likelihood of known predictors of good response to anti-TNF than traditional BASDAI definition (≥ 4) [13]. One study including patients with AS and undifferentiated SpA treated with etanercept or infliximab revealed that the discriminatory ability of the ASDAS outperformed that of the BASDAI, patient global score, ESR, CRP, or acute inflammation score as assessed by MRI [19]. In addition, a previous study investigating the levels and changes of 10 biomarkers in axial SpA patients during anti-TNF treatment revealed that the ASDAS-CRP was more strongly associated with these biomarkers than the BASDAI [20].

Moreover, the ASDAS seems to better predict the patient prognosis than the BASDAI. One prospective study of 2274 AS patients newly treated with anti-TNF agents indicated that a low ASDAS-CRP (< 2.1) result after 3 months of anti-TNF treatment predicted the risk of future non-disability better than a low BASDAI (< 4) result and low CRP concentration (< 5 mg/L) [14]. A 12-year longitudinal study using data from the OASIS cohort revealed that models including the ASDAS-CRP as the disease activity measure fitted the data better than models including the BASDAI, CRP, or both the BASDAI and CRP. In addition, an increase of one ASDAS-CRP unit led to an increase of 0.72 mSASSS units/2 years [21]. Similarly, in another study from the German SpA inception cohort, radiographic spinal progression was found to be associated with the time-averaged ASDAS-CRP (mSASSS progression by over 2 points/2 years: OR 1.80, 95% CI 1.04–3.13; syndesmophyte formation/progression: OR 2.45, 95% CI 1.26 to 4.77) [22]. However, no associations between radiographic progression and BASDAI results have been reported [22, 23].

Female sex was associated with high ASDAS-CRP results after 3 months (6 weeks for infliximab) of anti-TNF treatment, which may be attributed to the very small population of female patients in our study population: only five patients were female in the high-ASDAS group, while only 2 patients were female in the low-ASDAS group. On the other hand, several researchers have reported that females were less likely to respond to anti-TNF treatment and were more likely to discontinue anti-TNF treatment early based on data from registries of rheumatic disease including AS [24‐29]. However, to date, there has been no clear determination of the basis of sex-dependent differences in anti-TNF response and survival, and further study is therefore needed to clarify the relationships between anti-TNF treatment and sex.

The present study has some limitations. First, we only focused ASDAS-CRP although there are available novel activity indices for AS such as the BASDAS formula, similar to the ASDAS-CRP but uses only the BASDAI and CRP, and simplified ASDAS [30, 31]. Measuring ASDAS-CRP is not yet routine in clinical practice. Therefore, the size of our study population was relatively small and selection bias may have occurred due to the retrospective nature of the study; some patients suspected of high disease activity despite of low BASDAI might be assessed with ASDAS-CRP. Second, the number of female patients was small. Therefore, some results need to be interpreted with some caution. It might be explained by relatively high proportion of female patients having high BASDAI (≥ 4) after 3 months of anti-TNF treatment (12 females among total 61 patients) and consequently being excluded. One of possible explanations of it might be characteristics of female AS patients; several studies have shown that female AS patients tended to report high BASAI score compared to males [32‐35]. Or it might be associated to which we suggested in our study; females were less likely to respond to anti-TNF treatment. Third, we only considered disease activity when comparing survival during anti-TNF treatment, and did not consider the type of anti-TNF agent, doses, or dosing intervals in the analyses. Despite these limitations, the present study provides valuable real-world data obtained from AS patients treated anti-TNF agents and focused on discordances in disease activity that have rarely been assessed previously. We found a benefit of assessing the ASDAS-CRP with respect to prognosis and drug survival during treatment with anti-TNF agents. Although the applicability of ASDAS-CRP might vary across different regions since the measuring ASDAS-CRP needs laboratory tests and the calculator of digital system [31].

To conclude, a low BASDAI score after anti-TNF treatment may not be as meaningful as previously thought. About 40% of low-BASDAI patients may have high AS disease activity according to an ASDAS-CRP assessment and be at higher risk of discontinuation of anti-TNF treatment due to lack/loss of effectiveness. In addition, all cases in which anti-TNF treatment was discontinued due to clinical remission were in the low-ASDAS group. In consideration of our results and the great capacity of the ASDAS shown in previous studies, we suggest the use of the ASDAS-CRP alone or in addition to the BASDAI when assessing disease activity in AS patients treated with anti-TNF agents.