Study selection and baseline characteristics
Figure 1 shows a flow chart for trial selection. Our initial search yielded 3400 studies; of these, 383 were duplicates and 2947 were ineligible based on our screening of titles and abstracts. Thus, we retrieved full texts of 70 studies. Of these, 17 were not RCT, 9 were data missing, 2 were non adult patients with CKD, 11 were unrelated interventions or outcomes, 5 were unable to retrieve the original text, 6 were using the same data and 8 were without a non-exposed control group. In addition, 8 records were identified from citation searching, in which 2 were duplicates and 1 was without control group. Given significant heterogeneity in SUA level, only trials with SUA ≥ 7.0 mg/dl (420.0 μmol/L) in men or SUA ≥ 6.0 mg/dl (360.0 μmol/L) in women or at least mean baseline SUA ≥ 6.0 mg/dl (360.0 μmol/L) with no prior gout flares were included in this meta-analysis. Doria (2020) [15] and Momeni (2010) [24] were excluded for their baseline SUA level. The inclusion criteria for Doria (2020) [15] is SUA ≥ 267.75 μmol/L. The mean serum level of uric acid was 5.9 ± 1.2 mg/dL and 6.5 ± 2.2 mg/dL respectively in experimental and control groups in the population of the Momeni (2010) [19]. Meanwhile, Tanaka(2020) [25] was excluded because the study population included patients with prior gout. In the Golmohammadi (2017) study [29], as the researchers provided renal function-related data separately for CKD stage 3 and CKD stage 4, without offering data for the overall population, we consider this as two separate sub-studies: Golmohammadi-1(2017) and Golmohammadi-2 (2017), both of which are collectively included in the meta-analysis. Specific data from Mukri (2018) [30], including the lower quartile (Q1) and upper quartile (Q3) of eGFR, were unavailable. The absence of this information made it impossible to convert the reported “Mean (IQR)” (Interquartile Range) into “mean (SD)” (standard deviation). Nevertheless, data concerning acute kidney injury (AKI) from the Mukri (2018) [30] remain available for extraction and utilization in the meta-analysis.
Flow diagram of articles considered for inclusion. Annotation: abbreviation: CNKI, China National Knowledge Infrastructure
In total, 17 eligible studies with 2032 participants were included in the meta-analysis. Characteristics and demographic data from each of the 17 studies included in our review are listed in Table 1 [11,12,13,14,15,16,17, 28,29,30,31,32,33,34,35,36,37]. These studies were published between 2006 and 2023 and had sample sizes ranging from 40 to 441. There was no statistically significant difference between the ULT and control groups at baseline in most trails. In the 17 included studies, febuxostat dosage in the treatment group ranged from 10 to 80 mg/day and allopurinol dosage ranged from 100 to 400 mg/day. The control group in most trials was administered placebo or usual therapy or no treatment. Effects of febuxostat or allopurinol was assessed by measuring the changes in levels of SUA and Scr and the changes of eGFR, incidence of doubling of Scr without the requirement of dialysis and incidence of acute kidney injury (AKI).
Risk of bias
The methodologic quality of the results was evaluated by the Cochrane Collaboration risk-of-bias (ROB2) tool [26] and modified Jadad [22,23,24] scale. Eight (47%) of 17 trials were deemed of high quality, seven (42%) of 17 trials were deemed of moderate quality, two of (12%) of 17 trials were deemed of low quality (Supplementary Fig. 1). Six studies had a modified Jadad scale ranged from 1 to 3 which were considered as low quality, 11 trials had a Jadad scale ranged from 4 to 7 which were considered as high quality (Table 1). Four studies [11, 12, 26, 34] analyzed the intentionality of people who lost follow-up. Five randomized controlled trials were double-blind [11, 12, 14, 15]. Two trials [30, 34] were open label study. No crucial deviations from the intended interventions were reported in one of the 17 trials.
The change of the levels of uric acid
The levels of SUA were not significantly different at baseline between treatment and control groups in these 17 studies [11,12,13,14,15,16,17, 28,29,30,31,32,33,34,35,36,37]. Compared with the control group, ULT group lowered the level of serum uric acid with a weighted mean difference (WMD) of -160.54 μmol/L, 95% CI [-191.58, -129.51] (p < 0.001) with significant heterogeneity observed (I2 = 96.6%, p < 0.001) (Fig. 2). The result of the Egger’s test was statistically significant (p = 0.022) (Table 2), suggesting a risk of publication bias. However, the sensitivity analysis conducted by excluding individual studies demonstrated a relatively stable result (Supplementary Fig. 2).
Forest plot for the effect of ULT versus controls on thethe change in the level of uric acid. Annotation: controls, placebo or no treatment; ULT, uric acid-lowering therapy; the Golmohammadi (2017) [29] study were considered as two sub-studies: Golmohammadi-1(2017) and Golmohammadi-2 (2017); data are pooled WMDs with 95% CIs. WMD, Weight Mean differences; CI,confidence interval
Primary outcome: the change of eGFR
Ten RCTs [11,12,13,14,15,16,17, 29, 31, 34] involving 1521 participants reported the effects of ULT on the change in eGFR before and after interventions with the average of 12.7 months follow-up period. Overall, compared with the control group, ULT group preserved the loss of estimated eGFR by 3.67 mL/min/1.73m2, 95% CI[1.67,5.67], p < 0.001 with moderate heterogeneity observed (I2 = 48.2%, p = 0.037). The studies were.categorized into two subgroup based on their follow-up durations: short-term (3–6 months) and long-term (> 6 months). ULT preserved the loss of eGFR at short term (WMD, 5.74 mL/min/1.73m2, 95% CI[2.09,9.39]) and long term (2.07 mL/min/1.73m2, 95% CI[0.15,3.98]) (Fig. 3A), and the Egger’s test (p = 0.499, p = 0.096) showed low publication biases (Table 2).
Forest plot for the effect of ULT versus controls on the change in eGFR. A Categorized based on follow-up durations, B Subgroup-analysis according to renal function (the baseline mean eGFR). C Subgroup-analysis according to age. D Subgroup-analysis according to the countries of the included trails. Annotation: controls, placebo or no treatment; ULT, uric acid-lowering therapy; the studies were categorized into three segments based on their follow-up durations: short-term (3–6 months), long-term (> 6 months); the Golmohammadi (2017) [29] study were considered as two sub-studies: Golmohammadi-1(2017) and Golmohammadi-2 (2017); data are pooled WMDs with 95% CIs. WMD, Weight Mean differences; CI,confidence interval; eGFR, estimated glomerular filtration rate
What’s more, we analysed the change of eGFR data stratified by renal function (the baseline mean eGFR), subgroup analyses showed a significant renal benefit from ULT both in patients with the baseline mean eGFR ≥ 45 mL/min/1.73m2 (WMD 1.99 mL/min/1.73m2, 95% CI [0.21, 3.78] (p < 0.001)) and patients with the baseline mean eGFR < 45 mL/min/1.73m2 (WMD 2.00 mL/min/1.73m2, 95% CI [0.68,3.32] (p = 0.003)). The overall test for heterogeneity between two sub-groups was not significant (p = 0.376) (Fig. 3B).
We also performed a subgroup analysis by the baseline mean age of the included trails, subgroup analyses showed a significant benefit from ULT in patients with younger than 60 years old(WMD 4.76 mL/min/1.73m2, 95% CI [2.60, 7.00] (p < 0.001)), no significant heterogeneity was observed (I2 = 33.4%, p = 0.178), but not for patients more than or equal to 60 years old (WMD 1.08 mL/min/1.73m2, 95% CI [-0.54, 2.69] (p = 0.192), no significant heterogeneity was observed (I2 = 44.5%, p = 0.165).The overall test for heterogeneity between two sub-groups was significant (p = 0.008) (Fig. 3C).
Finally, we had intended to perform subanalysis by the countries of the included trails, however, there was a significant renal benefit from ULT both in patients from Asian countries (WMD 2.77 mL/min/1.73m2, 95% CI[1.19, 4.34] (p = 0.001)) and European and American countries (WMD 2.28 mL/min/1.73m2, 95% CI [0.29, 4.27] (p = 0.025)). The overall test for heterogeneity between two sub-groups showed no significance (p = 0.709) (Fig. 3D).
Sensitivity analysis was conducted by excluding individual studies demonstrated a relatively stable result analysis (Supplementary Fig. 3). What’s more, when analyzing only the RCTs of high quality (modified Jadad scale > = 4), 9 RCTs [11, 12, 14,15,16,17, 29, 31, 34] with high-quality were included in the analysis. There was also a significant renal benefit from ULT (WMD 3.53 mL/min/1.73m2, 95% CI [1.40, 5.65] (p < 0.001) (Supplementary Fig. 4A). Meanwhile, when deleting low-quality literature (assessed by ROB 2 tool), 8 RCTs [11, 12, 14,15,16,17, 29, 34] with high and moderate quality were included in the analysis, there was also a significant renal benefit from ULT (WMD 2.40 mL/min/1.73m2, 95% CI [0.85, 3.96] (p < 0.001) (Supplementary Fig. 4B).
Primary outcome: the change of levels of Scr
Nine RCTs [13, 28, 29, 31,32,33, 35,36,37] evaluated the change of levels of Scr (Fig. 4) in 728 CKD patients with asymptomatic hyperuricemia with the average of 10.2 months follow-up period. Overall, compared with the control group, ULT group reduced the increase of Scr (WMD -46.13, 95% CI [-65.64,-26.62]μmol/L (p < 0.001) with significant heterogeneity observed (I2 = 84.6%, p < 0.001). The study was.categorized into two subgroup based on their follow-up duration: short-term (3–6 months) and long-term (> 6 months). ULT reduced the increment of Scr both at short-term (WMD -44.48[-84.03,-4.92]μmol/L) and long-term (WMD -46.13 [-65.64,-26.62]μmol/L) (Fig. 4A). The Egger’s test (p = 0.075, p = 0.115) (Table 2), suggesting low risk of publication biases.
Forest plot for the effect of ULT versus controls on the change in Scr. Annotation: controls, placebo or no treatment; ULT, uric acid-lowering therapy; the studies were categorized into three segments based on their follow-up durations: short-term (3–6 months), long-term (> 6 months); the Golmohammadi (2017) [29] study were considered as two sub-studies: Golmohammadi-1(2017) and Golmohammadi-2 (2017); data are pooled WMDs with 95% CIs. WMD, Weight Mean differences; CI,confidence interval; Scr, serum creatinine
We conducted a subgroup analysis according to whether the baseline mean ages of participants more than or equal 60 or younger than 60 years old and subgroup analyses showed a significant benefit from ULT both in patients younger than 60 years old (WMD -32.55[-40.38,-24.73]μmol/L) (p < 0.001) and patients more than or equal to 60 years old (WMD -87.71 [-158.21,-17.21]μmol/L) (p = 0.015), the overall test for heterogeneity between two sub-groups showed no significance (p = 0.128) (Fig. 4B).
Nine RCTs [13, 28, 29, 31,32,33, 35,36,37] evaluated the change of levels of Scr are from Asian countries, so it is impossible to conduct a subgroup analysis between the European and American populations and the Asian populations.
Sensitivity analysis conducted by excluding individual studies one by one demonstrated a relatively stable result analysis (Supplementary Fig. 5). When analyzing with the high quality RCTs [29, 31, 36] (modified Jadad scale > = 4), there was also a significant renal benefit from ULT (WMD -26.91 μmol/L, 95% CI [-51.87, -1.95] (p = 0.001) (Supplementary Fig. 6A). When deleting low-quality literature (assessed by ROB 2 tool), there was also a significant renal benefit from ULT (WMD -55.08 mL/min/1.73m2, 95% CI [-83.65, -26.52] (p < 0.001) (Supplementary Fig. 6B).
Doubling of serum creatinine(Scr) without the requirement of dialysis
Five RCTs [28, 32, 33, 35, 37] were identified, providing data on the events of doubling of Scr without the requirement of dialysis for 355 patients. There were 22 (22 of 178, 12.4%) and 69 (69 of 177, 39.0%) events of doubling of Scr without the requirement of dialysis in the ULT and control groups, respectively. ULT significantly decreased the incidence of events of doubling of Scr without the requirement of dialysis (relative risk (RR) 32.0%, 95% CI [0.21, 0.49], p < 0.001) and no significant heterogeneity was observed (I2 = 0%, p = 0.653) (Supplementary Fig. 7). The Egger’s test (p = 0.077) suggesting a low risk of publication bias (Table 2).
AKI events
Three RCTs [13, 14, 30] were identified with AKI event (Supplementary Fig. 8), including 12 AKI events (12 of 274, 4.4%) in ULT group and 12 (12 of 268, 4.5%) in control group, who developed to AKI. There was no significant difference between the ULT and control groups (RR 97.0%, 95% CI [0.45, 2.12], p = 0.943), no significant heterogeneity was observed (I2 = 0%, p = 0.569). The Egger’s test (p = 0.638) suggesting a low risk of publication bias (Table 2).
- The Renal Warrior Project. Join Now
- Source: https://bmcnephrol.biomedcentral.com/articles/10.1186/s12882-024-03491-4