Case control study cohort
The salient demographic and clinical features of the present sample set and primary information of SPP1 polymorphisms analyzed in this study are presented in Additional files 2 and 3, respectively. For urolithiasis patients included in this study, the median age was 34 years with a gender distribution of 1.6:1 (Males: Females ratio), both of which were comparable with that of control group (p = 0.77 and 0.69 for age and gender distribution, respectively). 23% of patients presented at a younger age (<18 years). Almost half of the patients had multiple renal stones (41%), recurrent disease (49%), positive family history of urolithiasis (48%) and parental consanguinity (53%). SPP1 -593 T/A polymorphism was found to be monomorphic in this study. The allelic and genotypic distribution for SPP1 rs28357094:T>G and rs1126616:C>T SNPs deviated from the HWE in control group, and as a result, were not included in the subsequent data analyses. Sanger sequencing electropherograms for representative genotypes of each SPP1 polymorphism included in final analysis are presented in Additional file 4.
Data analysis for allelic and genotypic distribution suggested no significant association between the risk of urolithiasis and any of the SPP1 polymorphisms analyzed except for rs11439060:delG>G (OR = 0.40; p = 0.002 for G/dG genotype in co-dominant model) (Table 1). Additionally, SPP1 rs2853744:G>T polymorphism showed significant associated with increased risk of urolithiasis in a dominant model (OR = 3.14; p = 0.006). While, SPP1 rs11730582:T>C and rs11439060:delG>G polymorphisms were significantly associated with the risk of urolithiasis (OR = 1.78; p = 0.006 and OR = 1.60; p = 0.012, respectively) considering a recessive genetic model (Table 2).
Table 1: Association analysis of urolithiasis risk and SPP1 genetic variants considering their allelic and genotypic frequencies
SPP1 genetic variants
|
Genotype/Allele
|
Urolithiasis patients
n = 235, n (%)
|
Healthy controls
n = 243, n (%)
|
OR (95% CI)
|
p-value (corrected)†
|
rs2853744:G>T
|
T/T
|
07 (3.1%)
|
21 (9.1%)
|
Referent
|
0.024
|
G/T
|
61 (27.2%)
|
62 (26.8%)
|
2.95 (1.17-7.45)
|
G/G
|
156 (69.6%)
|
148 (64.1%)
|
3.16 (1.31-7.66)
|
T
|
75 (17%)
|
104 (23%)
|
Referent
|
0.035
|
G
|
373 (83%)
|
358 (77%)
|
1.44 (1.03-2.01)
|
rs11730582:T>C
|
T/T
|
63 (28%)
|
69 (29.6%)
|
Referent
|
0.017
|
T/C
|
87 (38.7%)
|
113 (48.5%)
|
0.84 (0.54-1.31)
|
C/C
|
75 (33.3%)
|
51 (21.9%)
|
1.61 (0.98-2.64)
|
T
|
213 (47%)
|
251 (54%)
|
Referent
|
0.056
|
C
|
237 (53%)
|
215 (46%)
|
1.29 (1.00-1.68)
|
rs11439060:delG>G
|
G/G
|
19 (8.3%)
|
12 (5%)
|
Referent
|
0.002
|
G/dG
|
65 (28.5%)
|
103 (43.3%)
|
0.40 (0.18-0.88)
|
dG/dG
|
144 (63.2%)
|
123 (51.7%)
|
0.74 (0.35-1.58)
|
G
|
103 (23%)
|
127 (27%)
|
Referent
|
0.170
|
dG
|
353 (77%)
|
349 (73%)
|
1.24 (0.92-1.68)
|
†p-values are corrected for age and gender. p-value adjustment for multiple testing using Bonferroni method was also made (p-value threshold 0.016). Statistical significance is highlighted in bold.
OR, odds ratio; n (%), frequency and CI, confidence interval.
Table 2: Association analysis of urolithiasis risk and SPP1 genetic variants considering dominant, recessive and log-additive genetic models
SPP1 polymorphisms
|
Model
|
Genotypes
|
Patients n = 235, n (%)
|
Controls n = 243, n (%)
|
OR (95% CI)
|
p-value (corrected)†
|
rs2853744:G>T
|
Dominant
|
T/T
|
07 (3.1%)
|
21 (9.1%)
|
1.00
|
0.006
|
G/T-G/G
|
217 (96.9%)
|
210 (90.9%)
|
3.14 (1.29-7.45)
|
Recessive
|
T/T-G/T
|
68 (30.4%)
|
83 (35.1%)
|
1.00
|
0.210
|
G/G
|
156 (69.4%)
|
148 (64.9%)
|
1.29 (0.87-1.90)
|
Log-additive
|
-
|
-
|
-
|
1.38 (1.01-1.89)S
|
0.040
|
rs11730582:T>C
|
Dominant
|
T/T
|
63 (28%)
|
69 (29.6%)
|
1.00
|
0.700
|
T/C-C/C
|
162 (72%)
|
164 (70.4%)
|
1.08 (0.72-1.62)
|
Recessive
|
T/T-T/C
|
150 (66.7%)
|
182 (78.1%)
|
1.00
|
0.006
|
C/C
|
75 (33.3%)
|
51 (21.9%)
|
1.78 (1.18-2.71)
|
Log-additive
|
-
|
-
|
-
|
1.26 (0.99-1.61)
|
0.062
|
rs11439060:delG>G
|
Dominant
|
G/G
|
19 (8.3%)
|
12 (5%)
|
1.00
|
0.150
|
dG/G-dG/dG
|
209 (91.7%)
|
226 (95%)
|
0.58 (0.28-1.23)
|
Recessive
|
G/G-dG/G
|
84 (36.8%)
|
115 (48.3%)
|
1.00
|
0.012
|
dG/dG
|
144 (63.2%)
|
123 (51.7%)
|
1.60 (1.11-2.60)
|
Log-additive
|
-
|
-
|
-
|
1.24 (0.92-1.67)
|
0.15
|
†p-values are corrected for age and gender. p-value adjustment for multiple testing using Bonferroni method was also made (p-value threshold 0.016). Statistical significance is highlighted in bold.
OR, odds ratio; n (%), frequency and CI, confidence interval.
Frequency of G-C-dG haplotype (SPP1 rs2853744-rs11730582-rs11439060 polymorphisms, respectively) was significantly higher in urolithiasis patients as compared to controls (OR = 1.68; p = 0.0079), suggesting an association with increased risk of urolithiasis in haplotype analysis (Table 3). However, pair wise linkage disequilibrium (LD) and haplotype plot structure analysis demonstrated lack of any substantial LD measures (based on D’ values) for each pair of SPP1 loci analyzed, suggesting that LD in this region is low (Additional file 5).
The SPP1 polymorphisms data was also analyzed considering sub-groups of Pakistani urolithiasis patients based on demographics (gender, age at first presentation), clinical features (stone multiplicity and stone recurrence) and histories (parental consanguinity and family history of urolithiasis). But, none of these comparisons yielded any significant associations (Additional file 6).
Table 3: Association of urolithiasis risk with of SPP1 genetic variants considering haplotype analysis
SPP1 haplotypes
(rs2853744:G>T- rs11730582:T>C- rs11439060:delG>G)
|
Haplotype frequency†
|
Case, control ratios
|
OR (95% CI)
|
p-value (corrected)‡
|
G-T-dG
|
0.323
|
0.313, 0.332
|
Referent
|
-
|
G-C-dG
|
0.312
|
0.374, 0.248
|
1.68 (1.15 - 2.46)
|
0.0079
|
G-C-G
|
0.090
|
0.086, 0.107
|
0.95 (0.57 - 1.58)
|
0.840
|
G-T-G
|
0.078
|
0.079, 0.068
|
1.25 (0.61 - 2.56)
|
0.550
|
T-T-dG
|
0.060
|
0.083, 0.053
|
1.43 (0.72 - 2.87)
|
0.310
|
T-C-dG
|
0.059
|
0.025, 0.081
|
0.53 (0.24 - 1.17)
|
0.120
|
OR – odds ratio; 95% CI – 95% confidence interval.
†Haplotypes with a frequency >5% were analyzed.
‡p-values are corrected for age and gender. p-value adjustment for multiple testing using Bonferroni method was also made (p-value threshold 0.0083). Statistical significance is highlighted in bold.
Meta-analysis
Qualitative synthesis for association of urolithiasis and SPP1 genetic variants
A flow diagram, reflecting the sequence of study selection for genetic association of 3 SPP1 genetic variants and susceptibility of urolithiasis, is described in Figure 1. An initial online search of literature databases by means of defined MeSH terms concerning urolithiasis and osteopontin genetic variants, resulted in retrieval of a total of 217 articles. However, in the end, a total of 4 studies were included in the present meta-analysis, comprising of 3 previously published reports obtained after rigorous screening according to the eligibility criteria, combined with the indigenous genetic epidemiology study. There were 2 studies exploring the association of SPP1 polymorphism rs2853744, 4 for rs11730582 and 3 for rs11439060. All included studies had NOS score of 6 or better.
The salient characteristics of the studies comprising the present meta-analysis are described in Table 4. The publication period for the selected studies ranged from 2010 to 2018. All four studies included in the meta-analysis were case-control studies, based on Asian population, and most (3/4) studies used control groups collected from general population. Also, all studies conformed to HWE with respect to their control group. Among these, three studies also analyzed the association of other polymorphisms in SPP1 gene with urolithiasis. However, data pertaining to additional polymorphisms were not included in the present meta-analysis. All studies showed least one of the analyzed SPP1 genetic variants to be positively associated with the susceptibility of urolithiasis.
Table 4: Main characteristics and findings of the eligible studies included in this meta-analysis.
Reference (first author, year)
|
Region
|
Ethnic group
|
Controls source
|
Samples (N)
|
Cases
|
Controls
|
Polymorphic sites
|
HWE status†
|
Genotyping method
|
Findings
|
Liu, 2010
|
Taiwan
|
Asian
|
Hospital based
|
496
|
249
|
247
|
rs11730582, rs11439060
|
Yes
|
TaqMan genotyping assay
|
rs11439060 of SPP1 promoter associated with risk of UL in allelic and genotypic models
|
Safarinejad, 2013
|
Iran
|
Asian
|
Population based
|
1026
|
342
|
684
|
rs2853744, rs11730582
|
Yes
|
PCR-FRET
|
SPP1 SNP rs2853744 showed significant association with UL
|
Xiao, 2016
|
China
|
Asian
|
Population based
|
480
|
230
|
250
|
rs11730582, rs11439060
|
Yes
|
TaqMan genotyping assay
|
rs11439060 in SPP1 promoter significantly associated with risk of UL as well as clinical characteristics in UL
|
Present study, 2018
|
Pakistan
|
Asian
|
Population based
|
478
|
235
|
243
|
rs2853744, rs11730582 and rs11439060
|
Yes
|
Sanger sequencing
|
All 3 SPP1 promoter SNPs associated with UL under different genetic models
|
†Yes indicates consistence with HWE.
FRET, fluorescence resonance energy transfer; HWE, Hardy-Weinberg equilibrium; N, Total number of samples; SPP1, osteopontin; SNP, single-nucleotide polymorphism; UL, urolithiasis.
Quantitative synthesis for association of urolithiasis and SPP1 genetic variants
For the association of SPP1 rs2853744 polymorphism, only 2 studies were available including 577 cases and 927 controls where overall results from recessive model reflected a statistically noteworthy association with the susceptibility of urolithiasis (OR = 1.37; p = 0.004, Figure 2b). However, no association was detected under dominant model after considering correction for multiple testing (Figure 2a). Measures of heterogeneity in this set of studies were not significant (I2 = 0%, p = 0.61 for recessive model) therefore, fixed effect model was employed to determine the cumulative results (Figures 2c and 2d).
Meta-analysis of SPP1 rs11730582 and rs11439060 polymorphisms included 4 (1056 cases and 1424 controls) and 3 (714 cases and 740 controls) studies, respectively. The summarization of all studies indicated no significant associations between rs11730582, and rs11439060 polymorphisms and urolithiasis using either a dominant or recessive model (Additional files 7 and 8, parts a and b) after correction for multiple testing. Heterogeneity analysis for SPP1 rs11730582 polymorphism was insignificant, but not for SPP1 rs11439060 polymorphism, therefor fixed and random effect models, respectively, were applied in calculation of pooled results (Additional files 7 and 8, parts c and d).
Shape of funnel plots and results of Egger’s test (as depicted in Figure 2, and Additional files 7 and 8) rendered calculation of the publication bias where absence of any significant publication bias was evident except for the analysis of SPP1 rs11439060 polymorphism and urolithiasis (Additional file 8, parts c and d); however, the Egger’s test was not significant (p = 0.19 for dominant model and p = 0.15 for recessive model) in that case too. The outcome in sensitivity analyses, performed by removing each study at a time, indicated the reliability of the current meta-analysis results as no significant influence of an individual study was evident on the cumulative OR.
The raw dataset containing the individual phenotypic and genotypic data for each of the SPP1 polymorphisms analyzed has been provided as additional file 9 in the supplementary data.