Estado de las masas de agua superficial

Just as peripheral joint assessment tools have been borrowed from RA, measures of axial disease activity have been used in PsA that were initially designed for use in AS. The most commonly used outcomes in AS are the Bath measures including the BASDAI, the Bath Ankylosing Spondylitis Functional Index (BASFI) and the Bath Ankylosing Spondylitis Metrology Index (BASMI). The BASDAI is a patient completed questionnaire with a series of six 100 mm VAS scales asking about fatigue, pain and early morning stiffness (Garrett et al. 1994). This is well validated in AS and has been used as the primary outcome in the vast majority of clinical trials in AS. The BASDAI provides a static measure of disease activity and cut points have been devised to define active disease. It is also routinely used as a response measure. Responses can require a reduction of a certain number of units or a reduction of a certain percentage such as a BASDAI50 which describes a reduction in BASDAI of at least 50%.

The BASDAI has been evaluated in a cohort of PsA patients to assess its use in this condition. This showed a high correlation between BASDAI scores and a self- reported VAS of current overall arthritis activity, suggesting that it is a good representation of patient’s perception of disease activity in PsA (Taylor and Harrison 2004). However this correlation was very similar in those with axial disease and those with peripheral disease suggesting that the BASDAI does not differentiate between axial and peripheral disease activity (Taylor and Harrison 2004). No other assessment tools of axial disease activity have been evaluated in PsA.

Related to the BASDAI is the BASFI, a measure of functional ability validated and used in AS (Calin et al. 1994). Like the BASDAI, this is based on a series of 0- 100 mm VAS scales where patients rate their ability to perform everyday tasks from easy to impossible. There are limited data concerning the use of the BASFI in PsA. One study of the BASFI showed a correlation with other functional indices in PsA patients but again it shows no difference in responses between patients with axial or peripheral disease (Leung et al. 2008). Therefore there is no obvious benefit to using the BASFI in preference over more established and validated tools such as the HAQ-DI and SF-36.

The BASMI is a combination of five clinical measurements of axial mobility developed from an assessment of twenty different measures in AS (Jenkinson et al. 1994). This shortened metrology index provides an assessment of spinal mobility including the cervical and lumbar spine. These measures have been assessed in PsA patients with axial disease using multiple observers. The International Spondyloarthritis Interobserver Reliability Exercise (INSPIRE) study showed that the intraclass correlation coefficients for patients with AS and axial PsA were similar and were excellent (ICC 0.89) (Gladman et al. 2007b). A further study tested the correlation between spinal mobility measures and radiographic evidence of spinal damage in both AS and PsA. This only used 19 patients (taken from the INSPIRE study) but did show a high correlation between the modified Stoke Ankylosing Spondylitis Spinal Score (mSASSS) and measures of spinal mobility (Chandran et al. 2007a). This provides some limited evidence of truth in the form of criterion validity.

Increasingly, composite measures including some of the Bath scales are being used to assess axial response in clinical trials in AS. The Assessment of SpondyloArthritis Society (ASAS) response measures were developed in 2001 using data from RCTs comparing NSAIDs versus placebo (Anderson et al. 2001). The authors initially included the five core outcomes established by the ASAS group (physical function, pain, spinal mobility, spinal stiffness/inflammation and the patient’s global assessment), but spinal mobility was excluded as there was no evidence of responsiveness to treatment. Multiple candidate definitions of improvement were tested looking for a low placebo response (<25%) and a significant difference between groups. The best criteria were the ASAS20 response criteria (see table 2) (Anderson et al. 2001).

These were then revised in 2004, following the advent of new therapies. TNF- inhibitors had proven efficacy in improving spinal mobility and acute phase reactants, which had been excluded from the previous response measures. When the original methodology was repeated using data from the RCTs of TNF inhibitors, the new ASAS20 (5 of 6 criteria) and a modification of the original ASAS response criteria using a 40% improvement were shown to be the best discriminators (see table 2) (Brandt et al. 2004).

ASAS criteria for response 2001 (Anderson et al. 2001)

Improvement of ≥20% and absolute improvement of ≥10 units in ≥3 of the following 4 domains:

•Patient global assessment (VAS global assessment scale)

•Pain (pain VAS)

•Function (BASFI)

•Inflammation (mean of questions 5&6 from the BASDAI)

with no deterioration, defined as a deterioration of ≥20% and net worsening of ≥10 units in the remaining domain.

ASAS response criteria 2004 (Brandt et al. 2004) ASAS 20 (5 of 6 criteria)

Improvement of ≥20% in five of the following six domains

ASAS 40

Improvement of ≥40% and absolute improvement of ≥20 units in ≥3 of the following 4 domains:

•Patient global assessment (VAS global assessment scale)

•Pain (pain VAS)

•Function (BASFI)

•Inflammation (mean of questions 5&6 from the BASDAI)

•Spinal mobility (BASMI)

•Acute phase reactant (CRP)

•Patient global assessment (VAS global assessment scale)

•Pain (pain VAS)

•Function (BASFI)

•Inflammation (mean of questions 5&6 from the BASDAI)

More recently, a new composite disease activity score has been developed by the ASAS group to provide a composite measure that could reflect disease activity in AS (Lukas et al. 2009). The Ankylosing Spondylitis Disease Activity Score (ASDAS) was developed from 12 candidate variables selected by ASAS experts. Data from 708 patients were used to develop the ASDAS. First, principal component analysis was used to identify sets of correlated variables to remove redundant outcome measures, then discriminant function analysis was used to investigate the contribution of the different factors in the separation of high and low disease activity. Finally, linear regression analysis was performed to provide a weighting for each of the discriminant variables. This analysis was used to produce four draft ASDAS formulae (see table 3) for future analysis. Cross validation was performed in a second cohort confirming similar discriminatory ability. The discriminatory ability, as measured by the standardised mean difference, was considerably higher for all four ASDAS when compared with any of the single variables (Lukas et al. 2009).

A= 0.122 x back pain + 0.061 x duration morning stiffness + 0.119 x patient global + 0.210 x √(ESR) + 0.383 x Ln(CRP+1) B= 0.079 x back pain + 0.069 x duration morning stiffness + 0.113 x patient global + 0.086 x peripheral pain/swelling + 0.293 x √(ESR) C= 0.121 x back pain + 0.058 x duration morning stiffness + 0.110 x patient global + 0.073 x peripheral pain/swelling + 0.579 x Ln(CRP+1) D= 0.152 x back pain + 0.069 x duration morning stiffness + 0.078 x fatigue + 0.224 x √(ESR) + 0.400 x Ln(CRP+1)

Further validation work has continued following this initial publication. Analyses in two further independent cohorts again showed that all four draft ASDAS were highly discriminatory (van der Heijde et al. 2009). Discussion within the ASAS group was used to decide on the final ASDAS formula to be recommended. It was decided that the ASDAS with CRP (ASDAS C in table 3) should be the preferred ASDAS score with an alternative ASDAS including ESR (ASDAS B in table 3) to be used if CRP is not available (van der Heijde et al. 2009). A further external validation study showed responsiveness to change and concurrent validity in patients with axial SpA treated with TNF blockers (Pedersen et al. 2010). Preliminary cut-points have also been proposed for different levels of disease activity and responses, although these need further investigation and validation before they are proven.

Unfortunately these composite measures (both ASAS response measures and the ASDAS) have not been assessed in PsA. Given the limitations of the BASDAI in PsA, there are justifiable concerns about their use; however the inclusion of other outcome measures may mean that these composite indices can be helpful in axial PsA. Ideally, studies with newer effective therapies are required in axial PsA to address this issue and to compare these different outcome measures.