Volumetric reduction and dissolution prediction of monosodium urate crystal during urate-lowering therapy – a study using dual-energy computed tomography

Abstract

Objective: Dissolution velocity of monosodium urate (MSU) crystal during urate-lowering therapy (ULT) had been inadequately studied. By using dual-energy computed tomography (DECT), which allows accurate assessment of MSU load, we analyze relationship between serum urate (SU) and volumetric reduction rate of MSU and develop a model that predicts dissolution time.

Methods

Baseline and follow-up DECTs were performed under a standard ULT protocol. Monthly dissolution rates were calculated by simple and compound methods. Correlations with average SU were compared and analyzed. Best-fit regression model was identified. MSU dissolution times were plotted against SU at different endpoints.

Results

In 29 tophaceous gout patients, MSU volume reduced from baseline 10.94 ± 10.59 cm³ to 2.87 ± 5.27 cm³ on follow-up (p = .00). Dissolution rate had a stronger correlation with SU if calculated by compound method (Pearson’s correlation coefficient r= −0.77, p = .00) and was independent of baseline MSU load. The ensuing dissolution model was logarithmic and explained real-life scenarios. When SU > 0.43 mmol/l, dissolution time approached infinity. It improved to 10–19 months at SU = 0.24 mmol/l. When SU approximated zero (as with pegloticase), dissolution flattened and still took 4–8 months.

Conclusion

MSU dissolution is better described as a logarithmic function of SU, which explains, predicts, and facilitates understanding of the dissolution process.

Keywords:

• Dual-energy computed tomography
• gouty tophi
• tophaceous gout
• urate-lowering therapy

Introduction

Chronic tophaceous gout is the end-stage disease of gout featuring macroscopic monosodium urate (MSU) crystallization in soft tissues and in joints [1]. It is associated with significant musculoskeletal morbidities such as joint destruction and mechanical blockade to tendon gliding, resulting in poor quality of life and functional impairment [2–4].

Tophus has long been known to reduce or resolve with urate-lowering therapy (ULT) [5,6]. Chemically, in vitro saturation threshold of urate is 0.40 mmol/l (6.8 mg/dl) at 37 °C [7]. Theoretically if serum urate (SU) is lower than the saturation point, crystal can be dissolved gradually. In a 10 years’ observation, SU was noted to relate with changes in tophaceous deposit detected physically [8]. Treatment target of SU ≤ 6 mg/dl (0.36 mmol/l) was first recommended for tophus reduction [8]. In a study in 2002, by serially measuring the diameters of superficial tophi, it was found that they shrank 0.5 mm per month at SU of 6 mg/dl (0.36 mmol/l) [9]. Tophus reduction velocity was for the first time shown inversely proportional to SU [9]. These findings led to the recommendation of treatment target of SU < 0.30 mmol/l (5 mg/dl) to facilitate faster tophus dissolution [10,11].

However, change in surface dimension of superficial tophus is an indirect assessment of MSU crystal burden. It does not necessarily correlate with the reduction of MSU load in the region [12,13]. In the calculation of tophus shrinkage, it assumed a constant rate of absolute dimensional change over time regardless of tophus size [9]. This may not reflect how MSU crystal dissolves in real-life. In a prospective pegloticase-treated series, it was observed that tophus loads of huge volume difference shrank by similar percentage over a comparable time frame [14].

Dual-energy computed tomography (DECT) is evolving as a specific radiographic assessment of MSU load, which should be the focus in delineating efficacy of ULT in the reduction of tophi. A meta-analysis confirmed high accuracy in diagnosing gout (sensitivity 0.85 and specificity 0.94) [15]. It allows detection of subclinical MSU deposit [16]. DECT was shown to be more reliable and reproducible in assessing tophus size and volume when compared to physical methods [13,17–19]. In assessing dissolution velocity, DECT should be the preferred method as it accurately measures the total MSU load in the region, while physical or ultrasonographic method represents sampling only [19]. Change in MSU load can be calculated by absolute terms or with reference to baseline. This should allow better visualization of the crystal dissolution process [11]. To our knowledge, there is no study investigating the relation between SU and the rate of MSU dissolution using DECT as assessment tool during ULT.

It is hypothesized that the rate of volumetric reduction of MSU load in gouty tophi on DECT correlates with SU during ULT. This in turns may provide a regression model that allows prediction of time required for tophus reduction.

Methods

A retrospective cohort study was carried out for patients with chronic tophaceous gout referred to the Department of Orthopaedics and Traumatology, New Territory West Cluster of the Hospital Authority of Hong Kong. We belong to a public hospital system with catchment area of 1.1 million people. Medications are charged at a nominal affordable fee and investigations are done free-of-charge.

The study was approved by our Cluster Research and Ethics Committee. The study period was from January 2015 to December 2019. Patients were managed under a standard protocol for medical dissolution of tophus in collaboration with the Departments of Medicine & Geriatrics and Radiology in our Cluster, according to the latest ACR and EULAR guidelines for gout (Appendix). As Han Chinese is the dominant ethnicity in Hong Kong, all patients were screened for HLA-B*58:01 and stratified to treatment with allopurinol or febuxostat as their key XOI [20]. Uricosuric agent like losartan may be added as adjunct [21]. A period of optimization of ULT for 3–6 months was allowed before baseline DECT was performed, which was the usual waiting time to arrange the scan. Patients were then followed-up at 4–6 months interval at the same dose of ULT. SU was checked on every visit. More frequent follow-up was arranged if drug compliance issue occurred.

For the baseline DECT, up to four regions with the most severe tophaceous involvement were selected for each patient. The regions were selected based on both the degree of external tophaceous involvement and the functional symptoms experienced by patients but not the frequency of gouty attacks. More region or total body scanning was not done to limit radiation exposure. Scans were performed using dual source dual energy CT scanner (Somatom Definition Flash, Siemens Healthcare, Erlangen, Germany). Parameters were 140 kV for one tube and 80 kV for the other. A 2-material decomposition algorithm was performed on a multi-technique CT workspace (Syngo.via VB10 software package, Siemens, Forchheim, Germany). MSU crystal was colour-coded as green and fused onto the standard greyscale CT image. Volume of MSU crystal (MSU load) of each region was measured. The minimal detectable volume was 0.01cm³, with minimum crystal diameter of 3mm [13,17,22].

A second DECT was done at 1 to 3 years after dissolution treatment or when tophi disappeared clinically, whichever came earlier. The original idea to include a second DECT in the clinical pathway was to document more accurately the degree of MSU dissolution for medication tail-down. Reference to the first set of DECT was made to allow direct comparison.

Subjects were recruited through the DECT registry of the Department of Radiology in our Cluster.

Inclusion criteria for the study were (i) patients with hyperuricaemia with gouty tophi proven by histology and/or DECT, (ii) on ULT optimized according to the clinical pathway for medical dissolution of tophus (Appendix), and (iii) two sets of DECT done. Exclusion criteria on patient basis were (i) low total MSU load on the baseline DECT (<0.2 cm³), (ii) treatment failure (SU persistently higher than 0.45 mmol/l), (iv) incomplete follow-up or follow-up too short (less than 8 months), and (v) tophus debulking surgery performed between the 2 DECTs.

Informed consents were obtained from patients fulfilling the above criteria to access their medical records for this study. Demographic data, SU before and during treatment, and MSU loads on DECT were collected. Clinical data like length of gout disease, previous use of ULT, status of superficial tophi and range of movement in fingers were recorded. Grip power was measured by Jamar Hydraulic Dynanometer (Summons Preston Inc., Bolingbrook, IL, USA) at position II for patients with hand and wrist involvement. Co-morbidities like hypertension and renal impairment and concomitant use of diuretics were noted.

The combined MSU load (summation of MSU volume of each scanned region) were analyzed. Assuming the MSU dissolution characteristics were similar in different regions and in different tissues, monthly dissolution rate of MSU load was calculated in different manners as follows: Absolute calculation – absolute volumetric dissolution rate (D_a) and cube root of D_a to simulate diameter reduction (D₃); percentage calculation – simple (linear) dissolution rate (D_s) and compound (logarithmic) dissolution rate (D_c). $$D_{\mathrm{a}}\mathrm{ }=\mathrm{ }\frac{A-B}{n}$$ $$D_3\mathrm{ }{=}^3\sqrt{\frac{A-B}{n}}$$ $$D_{\mathrm{s}}\mathrm{ }=\mathrm{ }\frac{\left(A-B\right)/A}{n}\times \mathrm{ }100\mathrm{\%}$$ $$D_{\mathrm{c}}\mathrm{ }=\mathrm{ }\left(1{-}^n\sqrt{\frac{B}{A}}\right) \times \mathrm{ }100\mathrm{\%}$$

(A is the baseline MSU load on DECT in volume; B is the follow-up MSU load on DECT in volume; n: number of months apart between A and B).

Statistical analysis was carried out by IBM SPSS software (version 23), except COCOR (http://comparingcorrelations.org/) was employed for comparison between correlations. Descriptive characteristics were reported in mean ± standard deviation (SD). Wilcoxon signed rank test was used to compare the difference between two non-parametric groups. Statistical significance was taken as p < .05. Correlations with average SU between the 2 DECTs were analyzed using Pearson correlation coefficient (r). Strong correlation exists if r ≥ 0.7, moderate if 0.5 ≤ r < 0.7, weak if 0.3 ≤ r < 0.5. Correlations were compared using Williams’ t test. Multiple regression analysis was performed with other independent variables including baseline MSU load and type of ULT. Linear regression with 95% confidence interval (CI) was performed in the scenario with SU or dissolution rate as dependent variable. Coefficient of determination (R²) was used to determine fitness of the straight line to represent the scattered data. Prediction of dissolution time at representative SU was made using 4 different endpoints (50%, 75%, 90% dissolution and 99% dissolution). Dissolution prediction curve for each endpoint was plotted on the same graph for better visualization. Both 90% and 99% could be regarded as treatment target of dissolution depending on baseline tophus load.

Results

Forty-six patients fulfilling inclusion criteria were identified. Seventeen were excluded (six for low MSU load, two for treatment failure, and eight for incomplete follow-up and one refused consent). Twenty-nine subjects were studied (all male). The average age was 62.69 years (range 33–87 years). Three had moderate severity and 26 severe degree of chronic tophaceous gouty arthropathy [10]. Twenty-Three patients had other medical comorbidities. The commonest medical comorbidities were cardiovascular diseases, for example, hypertension and ischemic heart disease. All patients had normal liver function in terms of serum alanine aminotransferace level. Patients’ baseline renal function was presented in Table 1. While patients presented with various degree of chronic renal disease, only end-stage renal failure requiring renal replacement therapy in one patient was classified as renal comorbidity.

Table 1. Patients’ age, change in MSU load, average serum urate and urate lowering therapy.

Patient number	Age at first DECT (Years)	Total MSU load in all scanned regions (cm^3)		Interval between DECTs^a (months)	Average serum urate level (mmol/l, ± S.D.^b)	ULT^c		Diuretics	No. of region scanned	Region^e	Baseline renal function (CKD^f stage)
		Baseline	Follow-up			XOI^d	Uricosuric
1	58	11.47	9.1	15	0.400 ± 0.146	Febuxostat 80mg/d	-	No	2	R foot, R elbow	3
2	53	5.21	0.65	35	0.363 ± 0.092	Febuxostat 80mg/d	-	No	2	R HW, L HW	1
3	43	0.65	0.14	29	0.374 ± 0.107	Febuxostat 80mg/d	-	No	1	R HW	1
4	69	15.72	1.47	33	0.340 ± 0.087	Allopurinol 200mg/d	Losartan 50mg/d	No	4	R HW, L HW, R elbow, L elbow	2
5	47	11.2	0.01	35	0.305 ± 0.050	Allopurinol 400mg/d	-	No	1	R knee	2
6	78	0.26	0.01	33	0.297 ± 0.099	Allopurinol 200mg/d	-	No	2	R HW, R elbow	3
7	74	7.39	0.05	32	0.310 ± 0.082	Allopurinol 400mg/d	Losartan 50mg/d	No	2	R HW, R elbow	3
8	53	25.2	0.01	32	0.256 ± 0.054	Febuxostat 80mg/d	-	No	1	R HW	2
9	69	1.88	0.07	28	0.330 ± 0.089	Allopurinol 400mg/d	-	No	1	L FA	2
10	83	38.65	19.81	13	0.380 ± 0.119	Febuxostat 80mg/d	-	No	4	R HW, L HW, R FA, L FA	1
11	66	41.32	19.14	14	0.368 ± 0.036	Allopurinol 600mg/d	-	No	2	R HW, L HW	1
12	57	10.61	0.95	23	0.306 ± 0.093	Febuxostat 80mg/d	-	No	4	R HW. L HW, R elbow, L elbow	1
13	62	2.18	0.02	25	0.275 ± 0.133	Allopurinol 400mg/d	-	No	2	R HW. L HW	3
14	55	0.58	0.04	21	0.327 ± 0.025	Allopurinol 300mg/d	Losartan 100mg/d	Yes	4	R HW, L HW, R elbow, L elbow	2
15	61	14.28	4.49	15	0.320 ± 0.054	Febuxostat 80mg/d	-	No	3	R HW, R FA, L FA	3
16	65	1.61	0.1	20	0.287 ± 0.055	Febuxostat 120mg/d	-	No	2	R HW, L HW	1
17	49	22.42	4.29	17	0.370 ± 0.061	Allopurinol 300mg/d	Losartan 100mg/d	No	2	R HW, L HW	1
18	54	18.23	2.4	13	0.275 ± 0.073	Allopurinol 400mg/d	-	No	4	R HW, L HW, R elbow, L elbow	3
19	77	10.2	2.4	15	0.250 ± 0.020	Allopurinol 400mg/d	-	No	1	L elbow	2
20	74	14.76	0.24	19	0.223 ± 0.020	Allopurinol 400mg/d	-	No	2	R HW, L HW	2
21	73	3.29	0.21	15	0.210 ± 0.221	Allopurinol 300mg/d	-	No	1	L HW	2
22	53	0.99	0.01	14	0.275 ± 0.032	Febuxostat 80mg/d	Losartan 50mg/d	No	1	L HW	3
23	75	5.42	0.51	12	0.290 ± 0.104	Febuxostat 80mg/d	-	No	1	R HW	3
24	56	0.88	0.05	12	0.280 ± 0.022	Allopurinol 400mg/d	-	No	1	R HW	1
25	64	3.64	0.24	11	0.253 ± 0.081	Febuxostat 40mg/d	-	No	2	R HW, L HW	2
26	61	1.92	0.04	11	0.290 ± 0.048	Allopurinol 100mg/d	Losartan 100mg/d	No	2	R HW, L HW	2
27	33	15.6	10.27	15	0.437 ± 0.029	Febuxostat 80mg/d	–	Yes	1	L HW	5
28	87	18.2	0.02	18	0.227 ± 0.107	Febuxostat 80mg/d	–	Yes	2	R FA, L FA	2
29	69	13.54	6.45	12	0.380 ± 0.102	Febuxostat 80mg/d	–	No	2	R HW, L HW	3

^aDECT: Dual-energy Computed Tomography; ^bS.D.: Standard Deviation; ^cULT: Urate lowering Therapy; ^dXOI: Xanthine Oxidase Inhibitor; ^eR: right, L: left, HW: hand and wrist, FA: foot and ankle; ^fCKD: chronic kidney disease.

The XOI used was either allopurinol 100mg to 600mg per day in 15 patients or febuxostat 40–120 mg per day in the rest (Table 1). Losartan was prescribed in six patients with dosage 50–100 mg per day. No probenecid or other uricosuric agent was used in this series. The average pre-treatment SU was 0.58 ± 0.11 mmol/l, while the average SU during treatment was 0.31 ± 0.06 mmol/l (p = .00). Twenty-one patients (72%) met the treatment target of 0.36 mmol/l and 14 patients (48%) met the desirable target of 0.30 mmol/l.

Thirty-eight DECT scans were performed in 24 patients over hands and wrists (14 were bilateral), 11 over elbows, 6 over feet and ankles and 1 over knee. In the hand and wrist region, the most frequently involved structures were carpal tunnels and flexor tendons, and these were closely related to patient’s functional symptoms (Figure 1).

Figure 1. Distribution of MSU crystal before and after treatment. The paired foot and ankle DECT images belong to Patient 28. MSU load in the left foot reduced 99.9% from 8.36 cm³ to 0.02 cm³. The paired hand and wrist DECT images belong to patient 17. MSU load in the right hand reduced 86% from 9.01 cm³ to 1.28 cm³. Flexion contracture in middle finger improved. Power grip improved 2.8 times. (Remark: Green colour represents MSU crystal).

At baseline, the average combined MSU load per patient was 10.94 ± 10.59 cm³. In the second scan it was reduced to 2.87 ± 5.27 cm³ (p = .00). The time between two sets of DECT was 20.2 ± 8.0 months. Clinically, all patients exhibited significant softening or reduction of their superficial tophi. Tophi completely disappeared clinically in 22 patients (76%).

Correlation

No correlation was shown between SU and the absolute rate of volumetric change D_a (r = −0.08) or its cube root D₃ which simulated diameter change (r = 0.01). Simple monthly dissolution rate D_s showed a moderate correlation with SU (r = −0.58, p = .00) (Figure 2(a)). Under compound depreciation model, monthly dissolution rate D_c showed a strong correlation with SU (r = −0.77, p = .00) (Figure 2(b)). The correlation between D_c and SU is significantly stronger in negative correlation than that of D_s (p = .038, one-sided). Logarithmic rather than simple linear relationship between SU and MSU dissolution was favored.

Figure 2. (a and b) Serum urate as a function of monthly tophi dissolution rate drawn in comparison to Perez-Ruiz et al. [9] (a) Dissolution graph of combined MSU load at simple rate; (b) Dissolution graph of combined MSU load at compound rate).

Multiple regression analysis showed that D_c was independent of patient’s age, disease duration, baseline MSU load, standard deviation of SU, the use of allopurinol or febuxostat, or concomitant administration of diuretics. The additional use of uricosuric (losartan in this series) increased D_c by 3.6% (95% C.I. 3.1–11.9%, p = .00). It is also independent of various co-morbidities of patients and the scanned regions, except for elbow. Patient with tophaceous involvement of elbows which were scanned showed 3.6% reduction of D_c (95% C.I. 2.8–10.9%, p = .00).

Prediction and projection

Since dissolution rate had a stronger correlation with SU if calculated by compound method, the relation between SU and volumetric reduction of MSU is better described as logarithmic. Their regression model was chosen for projection and prediction.

With SU as the independent variable, compound monthly dissolution rate y% can be calculated (or plotted from Figure 3(a)) at a given SU level × by their linear regression: $$y=47.4\mathrm{ }–\mathrm{ }109.3x$$

Figure 3. Prediction of MSU (tophi) reduction rate and time with serum urate level. (a) Compound monthly dissolution rates as a function of serum urate level between 0.18 to 0.48 mmol/l (Solid line – linear regression line, dotted lines – the limits of 95% confidence interval). At 0.36 mmol/l, the compound monthly dissolution rate was 8.5% on the average, at 0.30 mmol/l around 14%, and 0.24 mmol/l around 20%. (b) MSU dissolution model helps us to understand the dissolution process. It is dissolution time as a function of serum urate for representative dissolution percentages (t99 = time for 99% MSU dissolution, t90 for 90%, t75 for 75% and t50 for 50%). The upper end of the curves reflects situation with failed ULT. Some degree of dissolution may still be possible up to 0.43 mmol/l (7.2 mg/dl). Complete dissolution is impractical if SU > 0.36 mmol (6 mg/dl). By extrapolating our data towards zero SU, the lower end reflects cases responding to pegloticase. It still takes 4–8 months to dissolve most MSU crystal when SU approaches zero. At each SU level, dissolution process is conceptualized as one migrates from time zero to t50 then to t99. Dissolution time gets longer and longer as one approaches complete dissolution.

At SUs of 0.36 mmol/l (6 mg/dl), 0.3 mmol/l (5 mg/dl) and 0.24 mmol/l (4 mg/dl), the projected monthly reduction rates of gouty tophi were 8.1%, 14.6% and 21.2% (95% CI 5.4–10.8%, 12.6–16.7%, 18.0–24.5%), respectively (Figure 3(a)).

To predict time required for dissolution (tz) where z is the % tophus dissolution, the formula for compound depreciation was employed: $$tz=\frac{\text{log}[\frac{100-\text{z}}{100}]}{\log \left[\frac{100-y}{100}\right]}$$

At SU of 0.36 mmol/l (6 mg/dl), 90% MSU crystal dissolves in 27.3 months (95% CI 20.1–41.5 months). At SU of 0.30 mml/l (5 mg/dl), 90% MSU crystal dissolves in 14.6 months (12.6–17.1). At SU of 0.24 mmol/l (4 mg/dl), it further shortens to 9.7 months (8.2–11.6).

To better predict dissolution time and to visualize the dissolution process, we combine the above equations by substituting y with x. Estimated time for 50%, 75%, 90% and 99% dissolution of MSU crystal (t50, t75, t90 and t99) in terms of months was plotted against SU level (mmol/l) as follows in Figure 3(b): $$t50=\frac{-0.3}{\log \left(0.53+1.09x\right)}$$ $$t75=\frac{-0.6}{\log \left(0.53+1.09x\right)}$$ $$t90\mathrm{ }=\mathrm{ }\frac{-1}{\log \left(0.53+1.09x\right)}$$ $$t99\mathrm{ }=\mathrm{ }\frac{-2}{\log \left(0.53+1.09x\right)}$$

The trend of each curve (tz) was the dissolution time as a function of SU for a specific dissolution percentage (z). It was logarithmic. As SU increased from 0.36 to 0.43 mmol/l, the dissolution time rose exponentially. The curves gradually flattened towards the lower range of SU. The dissolution times did not reduce much as SU dropped from 0.18 mmol/l to zero.

At each SU level, dissolution process was represented when one migrated from baseline (t0) to t50 then to t99. Comparing 50% and 75% dissolution, time was doubled; from 90% to 99%, it was doubled too. The change was again logarithmic with dissolution time got longer and longer as one approached complete dissolution.

Discussion

The present study challenged the traditional concept in tophus dissolution, that is, tophi are expected to reduce the same dimension every month at a given SU level. It is generally quoted that at SU 6 mg/dl (0.36 mmol/l) tophi shrink by 0.5 mm/month or 1.5 mm/month at level ≤4mg/dl (0.24 mmol) [23,24]. These figures imply small tophi shrink very fast, but large tophi shrink slowly at least in the beginning following reduction of SU. Instead we showed that volumetric dissolution rate of MSU was independent of the baseline load. That is, large MSU load took similar time to dissolve a same percentage when compared to small MSU load at the same SU level. Besides our own data as listed in Table 1, such finding was also observed in a prospective pegloticase-treated series using DECT as monitoring tool [14]. However, this should not be mixed up with the time to achieve complete resolution in other studies. Large tophus needs longer time to shrink to the same negligible size compared to small tophus. Therefore, we presented two treatment targets (t90 and t99) in the dissolution prediction for small or large MSU load.

DECT scan was performed on regions that were most severely affected clinically regardless of frequency of gouty attack. In our series, patients had more upper limb functional complaints than lower limb, accounting for more hands and wrists scanned than feet. Apart from superficial tophi that were easily picked up, significant portions of deep tophi at the carpal tunnel and flexor tendons were identified by DECT. While superficial tophi represent biased samples of the total MSU load in a region, DECT measures the total MSU volume in the region, part of which may not present as tophi [12,22]. In the present series, 22 patients (73%) had complete resolution of superficial tophi. On DECT, 18 patients (60%) achieved 90% MSU dissolution while only 5 patients (17%) achieved 99% dissolution. In a study with patients well-managed by ULT (SU < 6.0 mg/dl) without tophi, 47% had MSU crystal detected on DECT with volumes ranging from 0.01 to 0.89 cm³ [16]. In another study looking at gout patients fulfilling remission criteria and no tophus, MSU crystal of up to 1.23 cm³ was detected in 64% of them [25]. It was suggested that tophus size correlated poorly with the true MSU load. On the other hand, tophus contains both urate and non-urate (soft tissue) portions [22,26]. The MSU deposition is directly affecting the soft tissue volume of tophus [22]. They in turn have independent effects on bone and joint erosion [22,27]. Therefore, monitoring of MSU dissolution by DECT is more relevant in the treatment of tophaceous gout than tophus measurement [28].

In the MSU dissolution process with ULT, we found that compound (logarithmic) dissolution rate calculation was a better model to describe the relationship between SU and volumetric reduction of the crystal. This was illustrated in the dissolution model in Figure 3(b). While we could not validate the dissolution in the whole range of SU by statistics, the dissolution curves that we extrapolate from our data explain well real-life situations at both extreme ends of SU. When SU approaches its saturation point, it is imperative that dissolution time would get exponentially long. On the lower extreme, the availability of pegloticase allows examination of the dissolution process when SU approaches zero. In a meta-analysis of two randomized controlled trials on tophus reduction using pegloticase, around 35% recruited patients were responders to pegloticase [29]. Among these responders, SU was reduced to an average of 0.029 mmol/l (0.49 mg/dl). Complete resolution of all tophi was achieved in 35% of responders in 6 months’ time and the mean time to resolution of all tophi was expected to be 9.9 months. This is in line with our dissolution model that it would take around 4–8 months for 90–99% MSU dissolution.

The dissolution model (Figure 3(b)) represents the first attempt in literature to estimate duration for tophus resolution at a given SU level. Curves of different dissolution percentage were plotted on the same graph to facilitate understanding of the dissolution process. At the upper therapeutic range of SU up to 0.43 mmol/l (7.2 mg/dl), some degree of dissolution is still possible within a reasonable time frame. This is consistent with the finding that tophi could be partly resolved on DECT over 22 ± 11 months through lifestyle change alone when SU was brought down to an average of 6.7 ± 1.7 mg/dl [30]. However, complete dissolution is impractical if SU > 0.36 mmol (6 mg/dl). When SU was lowered from 0.36 mmol/l (6 mg/dl) to 0.24 mmol/l (4 mg/dl), 90% dissolution would be shortened from 27 months to 10 months. This substantial reduction provides a clear incentive to physicians and patients to maximize XOI and add uricosuric.

While lower SU hastens dissolution, maintenance of super-low level (below 0.18 mmol/l or 3 mg/dl) is not recommended [11]. There may be potential neuro-protective effects of urate [31]. Febuxostat, being a strong XOI capable of achieving low SU [32], is also associated with excessive cardiovascular mortality when compared to allopurinol on the long-term [33]. After a period of aggressive treatment, ULT is stepped down for cost and side effects consideration [34]. SU of 0.38 mmol/l (6.4 mg/dl) would be a reasonable maintenance target (Figure 2(b)).

Achieving complete resolution of tophus is well known to be a difficult task [35,36]. The concept of treat-to-target has been advocated for chronic tophaceous gout for more than 10 years [37,38]. Yet, in a 5-year cohort study published in 2020, none of the severe gout patients achieved remission [35]. Poor drug compliance is often blamed for [39,40]. The 2012 American College of Rheumatology guidelines had warned that significant pitfalls in patient education and adherence have been identified in gout [10,40]. Discrepancies in both the understanding of rationale of ULT and the perception of medical adherence between physicians and patients were noted [41]. The new findings in the present study provided not only scientific support to the desirable treatment target in advanced gout but also important information to patients on treatment expectation to enhance adherence [35,42].

The main drawback of the current study is the retrospective design. The compound dissolution rate calculation and the logarithmic model of dissolution could be over-simplified. The lack of data in the lower therapeutic range of SU limited the prediction accuracy in that region. Prediction would be most accurate in the middle range of SU (0.24–0.36 mmol/l) (Figure 3(a)). Even so, it was well observed that different bodily regions had different dissolution pattern. Our regression analysis suggested elbow tophi may be slower in dissolution and the prediction model could be more relevant to tophi in hand and wrist or foot and ankle. Dissolution also varies in individuals due to fluctuation in SU during ULT and other factors affecting crystallization [43]. Another possible source of error is the time delay in obtaining DECT. Due to resource constraint, it usually took 3–6 months to arrange DECT appointment. For patients with low MSU load on second DECT, some of the tophi may have completely dissolved before the follow-up scan, resulting in underestimation of the dissolution rate. While the prediction is for reference only, it put our understanding of MSU crystal dissolution into perspective. We hope this dissolution model could be further improved and better validated with more data added in the future by prospective study.

In conclusion, with the help of DECT, the present study shed new insights on the dissolution of MSU crystal in tophaceous gout in clinical practice. The compound monthly dissolution rate correlated strongly with SU and was independent of the baseline MSU load. A novel dissolution prediction model was presented with reference to SU. Logarithmic pattern of dissolution explains observations in real life. The prediction illustrates the benefit of tightening treatment target. It also conceptualizes the dissolution process at each SU level.

Ethics approval

This research was approved by New Territories West Cluster Research Ethics Committee (EC ref: 19082). Informed consent was obtained for all recruited subjects.

Acknowledgements

The authors thank Dr. Ming Keung Yuen, retired Consultant Radiologist, Tuen Mun Hospital, Hospital Authority, Hong Kong for his inspiration and kind support to this study and Mr. Jaden Lam, Statistical Officer, New Territories West Cluster Research Office, Hospital Authority, Hong Kong, for his excellent statistical support to the data analysis.

Conflict of interest

None

Additional information

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Appendix:

Clinical pathway for medical dissolution of tophus