Monosodium urate burden assessed with dual-energy computed tomography predicts the risk of flares in gout: a 12-month observational study

Gout is a metabolic condition related to monosodium urate (MSU) crystal deposition in joints and soft tissue leading to NLRP inflammasome-guided recurrent arthritis [1]. Despite the permanent deposition of MSU crystals, flares of inflammatory response are only intercurrent and difficult to predict. Risk factors for gout flares are known or suspected but causality is poorly understood [2, 3].

Initiation and modifications of urate lowering therapy (ULT) are critical times for flares. Prophylaxis of flares with colchicine, non-steroidal anti-inflammatory drugs (NSAIDs) or even oral corticosteroids is recommended by all international guidelines during 6 months following ULT initiation [4‐6]. These recommendations are based upon the observation of an increased rate of flares during the first 6 months of ULT in most randomized controlled trials (RCTs) involving ULTs versus placebo [7, 8]. Prophylaxis, especially with colchicine, has been proven effective to decrease the risk of flares [9‐11]. It is generally accepted that the reduced serum urate (SU) [12] concentration induces mobilization of the deposited MSU burden, potentially exposing the crystals to the innate immune system [11, 13]. This explanation, however, remains hypothetical.

Ultrasonography (US) and dual-energy computed tomography (DECT) can provide an assessment of the MSU burden [14]. DECT uses two x-ray beams with two different energies allowing to distinguish between urate and calcium in soft tissues surrounding bone, with radiation exposure close to conventional computed tomography (CT) [15]. US, on the other hand, can identify intra-articular cartilage MSU deposition appearing as a double contour (DC) sign, which disappears during urate depletion [16]. The usefulness of US and DECT is now fully recognized for the diagnosis of gout [17]. The clinical relevance of observing and measuring the urate burden by any imaging technique however needs to be determined. Previous data suggest an association between urate burden and past flares [18] but to our knowledge no imaging feature has been associated with the risk of flares [19]. More generally, no data have so far demonstrated that imaging adds to the low cost clinical and biological assessments in the management of gout [19].

The objective of this study was to determine if the extent of urate burden measured with DECT and US predicts the risk of gout flares.

This study demonstrates that the extent of the MSU burden measured with DECT but not US predicts the risk of flares. Interestingly, patients with dramatically large volumes of MSU deposition are not necessarily at risk of flare. This is the first study showing the usefulness of DECT for the management of patients with gout, beyond diagnosis. These results also provide proof of concept that urate load is associated with the risk of flares.

With an OR of 2.03, the MSU burden in the feet assessed with DECT is the strongest determinant of subsequent gout flares identified so far. Some comorbidities, namely hypertension, renal disease and coronary heart disease, have been identified in a population-based study as having hazard ratios of 1.1–1.3 of incident gout flares [3]. Of note, patients with high blood pressure (HBP) were less at risk of flare in our study (Table 1). SU change from baseline has been shown to have an impact on the risk of flares in the febuxostat trials, but the OR was close to 1. This is consistent with our results as changes in SU levels between M0 and M6 were similar between participants experiencing and not experiencing flares. Rapidly obtaining SU levels below the dissolution threshold of MSU was associated with an OR of 1.42 of incident flares in the febuxostat trials and could be related to the mobilization of the MSU burden [13]. Prevalence of the US DC sign provides a different assessment of MSU burden than the one provided with DECT [14] and does not seem to predict the risk of flares. Flare prophylaxis is known to imperfectly prevent gout flares during ULT initiation [11]. The variety of situations included in our study, both in terms of ULT (absence, dose adaptation, initiation) and types of prophylaxis, would probably explain why it did not exhibit a preventive effect.

When the question of introducing or pursuing flare prophylaxis arises, DECT could be a decisive partner in the decision-making process. So far, international societies recommend flare prophylaxis during 6 months or until tophus resolution [4‐6]. Becker et al. had shown from the febuxostat trials that the presence of subcutaneous tophi was one of the determinants of flare incidence during follow up [13]. Physical examination of tophi can be sufficient to show the presence of a significant MSU burden requiring flare prophylaxis. However, only approximatively 15–30% of patients present with subcutaneous tophi [24, 25]. Our study shows that DECT can provide an assessment of the risk of flares whether subcutaneous tophi are present or not. DECT scanning could be decisive when considering the interruption of flare prophylaxis after 6 months or when tophi are no longer clinically detectable.

MSU burden could be the missing link between ULT and flare reduction. So far, SU has been the primary endpoint in the majority of ULT trials, as observational data consistently show that eventually flare reduction correlates with reduction in SU [26]. In its latest highly debated guidelines, the American College of Physicians has challenged the relevance of targeting SU rather than flares themselves, considering that the evidence linking SU and flares is insufficient in RCTs [26‐29]. For instance, the double blind RCTs for the development of febuxostat failed to show flare reduction with febuxostat [30, 31]. Similarly, another RCT of febuxostat versus placebo in early gout showed no reduction after 6 months of treatment [8]. Given that it has already been shown that ULT in its most potent form (pegloticase) decreases the MSU burden assessed with DECT and that our study demonstrates that MSU volumes measured with DECT predict gout flares, DECT could be used as a clinically relevant outcome measure [32]. Depending on the kinetics of MSU depletion with conventional ULTs (hypoxanthine oxidase inhibitors and uricosurics), DECT MSU volumes could be considered as a surrogate marker for the risk of flares potentially more efficiently than SU.

We acknowledge that the study design had some limitations, imposing caution while interpreting the results. First, although an effect of volume of the MSU burden measured by DECT on the risk of flares was detected despite the small sample size, the rather small number of patients and the missing data may have underestimated other factors with a smaller effect. Particularly, the study did not detect the expected significant effect of the change in SU levels on the risk of flares [33]. However, the SU change between patients with and without flare was very similar in our study, suggesting that the link between change in SU level itself and the risk of flares may not be as direct as previously hypothesized. For instance, the high risk of flares during pegloticase therapy (and to a lesser extent during less potent ULTs) could be more the reflection of rapid depletion in MSU burden than SU change itself [32, 34, 35]. Second, the study included a panel of treatment initiators, patients already treated needing reinforced therapy and patients remaining under stable ULT treatment, which does not allow us to determine if flare prediction by DECT could be applicable at all stages in gout management. Future studies should assess the relationships operating in each of these subpopulations. Third, the specificity and sensitivity of the 0.81 cm³ cutoff for MSU deposits in the feet above which there is a significant risk of flaring were probably over-estimated, given the fact that they were calculated based upon the sample that was also used to build the model. Specificity and sensitivity testing in another population is needed. Fourth, 14 patients were lost to follow up which is expected in a population with gout who are known to have compliance issues. In addition, patients did not attend all visits, which may have led to a selection bias towards patients who experienced flare. Fifth, definition of gout flares in this study relied on patient’s self-assessment and retrospective confirmation by the physician. Although gout flares are marking painful experiences, the recently published definition of flares was not applied as such and the recollection of having experienced a flare may have been wrong for some patients [21, 36]. Sixth, the accuracy of MSU volume measurement with DECT is still debated, but measurements were standardized and known artifacts removed [37‐39]. Finally, patients were assessed at each visit but there was no assessment of treatment adherence (prophylaxis and ULT) between visits that could potentially have an unmeasured impact on the risk of flares. The same can be said for known triggers for gout attacks such as alcohol intake or intercurrent infection. It can only be assumed that patient behavior was similar between groups.

We believe this study is a step forward in our ability to tailor gout management to individual patients’ initial profiles. Particularly, special caution should be given in the prevention of flares in patients with initial high urate burden.