Review Article
Association between blood lipid profile and urolithiasis: A
systematic review and meta-analysis of observational studies
Huseyin Besiroglu
1and Emin Ozbek
21
Department of Urology, Catalca Ilyas Cokay State Hospital, and2Department of Urology, Cerrahpasa Medicine Faculty, Istanbul University, Istanbul, Turkey
Abbreviations & Acronyms BMI = body mass index CI = confidence interval CT = computed tomography GFR = glomerularfiltration rate HDL = high-density lipoprotein HR = hazard ratio LDL = low-density lipoprotein
Mets = metabolic syndrome NOS = Newcastle–Ottawa Scale OR = odds ratio TC = total cholesterol TG = triglyceride USG = ultrasonography Correspondence:Huseyin Besiroglu M.D., Department of Urology, Catalca Ilyas Cokay State Hospital, Istanbul, Turkey. Email:
drhuseyin1985@hotmail.com Received 2 February 2018; accepted 8 July 2018. Online publication 27 August 2018
Abstract: The objective of this study was to pool individual studies regarding the association of blood lipid profiles with urolithiasis to carry out a systematic review and meta-analysis. We searched MEDLINE, PubMed, Embase and Cochrane Library to identify the relevant studies up to November 2017. Studies that met all inclusion criteria were chosen, and a pooled analysis of the odds ratio between urolithiasis and dyslipidemia traits was calculated. A total of 11 observational studies (seven cross-sectional, three cohort, one case–control) with a total of 282 479 participants were examined. The overall pooled analysis of eight studies showed that high triglyceride was associated with increased estimated risk of urolithiasis (odds ratio 1.287, 95% CI 1.073–1.544; P = 0.007). Estimates of the total effect size were consistent in the sensitivity analysis. No evidence of publication bias was detected. The overall pooled analysis of nine studies showed low high-density lipoprotein was weakly associated with increased estimated risk of urolithiasis (odds ratio 1.171, 95% CI 1.010–1.358; P = 0.032). The sensitivity analysis showed conflicting results. No evidence of publication bias was detected. Three studies on the association between any dyslipidemia traits and urolithiasis showed a significant association (odds ratio 1.309, 95% CI 1.202–1.425; P < 0.001). The present meta-analysis showed that patients with higher triglyceride and lower high-density lipoprotein had an increased estimated risk of urolithiasis. A triglyceride–urolithiasis association was found to be more coherent and consistent compared with the high-density lipoprotein–urolithiasis association. Although somewhat contradictory results have been found, the meta-analysis is encouraging for evaluating urolithiasis as a systemic disorder. Further well-designed prospective randomized controlled or cohort studies are necessary to better elucidate the causal association of dyslipidemia and urolithiasis.
Key words: dyslipidemia, high-density lipoprotein, metabolic syndrome, triglyceride,
urolithiasis.
Introduction
Urolithiasis is a common disorder with an increasing prevalence rate. It is estimated that 5– 15% of the USA population develops the symptomatic stone disease by the age of 70 years.1–3 Despite its high prevalence and recurrence rate, the exact stone formation mechanisms are not yet elucidated. Many studies have shown that the prevalence of chronic diseases, including dia-betes mellitus, obesity and hypertension, is high among stone formers compared with healthy counterparts.4–10 Mets, a constellation of interrelated conditions including central obesity, impaired fasting glucose, dyslipidemia and hypertension, has raised keen interest for many dis-eases. Several studies have shown a positive association between Mets and urolithiasis.11,12 Dyslipidemia is an integral component of Mets, and current evidence suggests that dyslipi-demic patients are more likely to have a higher prevalence of uric acid and calcium oxalate dihydrate calculi, as well as a lower urinary pH.13,14
Although it has not yet been clearly defined, the background and pathophysiology of the association between the lipid profile and nephrolithiasis include inflammation and oxidative stress–anti-oxidant arrangements. Schmiedl et al. reported a relationship between a diet rich in fat and urinary abnormalities in Sprague–Dawley rats fed a cholesterol- and fat-rich diet
for 111 days compared with rats fed with standard laboratory chow.15 Another animal model study by Fujii et al. showed that adiponectin inhibited the kidney crystal formation in Mets model mice through inhibition of inflammation and apoptosis.16 These results are encouraging and helpful for designing the clinical studies evaluating the implication of lipid metabolism disturbances in nephrolithiasis, as well as finding new possible targets to prevent stone recurrence.
Either alone or as part of Mets, dyslipidemia and its possi-ble association with urolithiasis was evaluated in some indi-vidual epidemiological studies, but they are not sufficient to arrive at a precise and coherent conclusion. The current scien-tific evidence has presented conflicting results on the associa-tion of dyslipidemia traits (HDL, LDL, TC) with urolithiasis. Accordingly, we decided to combine them in a systematic review and a meta-analysis design with the aim of shedding some light on a debate to determine whether there is an asso-ciation between urolithiasis and serum lipid profile.
Methods
Literature search
The PubMed, MEDLINE, Embase and Cochrane databases were independently searched by two investigators to retrieve relevant studies published before October 2017. Discrepancies were resolved by consensus. The search was restricted to English language articles and studies of human participants. The search terms comprised of the following keywords: urolithiasis, nephrolithiasis, renal stone, renal calculus, renal calculi, kidney stone, kidney calculus, kidney calculi, dyslipi-demia, cholesterol, TG, HDL, LDL and Mets.
Study inclusion and evaluation
Studies were included in the present meta-analysis if they met the following criteria: (i) the study design was observa-tional; (ii) the outcome of interest was the prevalence or inci-dence of urolithiasis in patients with dyslipidemia; and (iii) OR or HR and corresponding 95% CIs (or data to calculate them) were reported.
The exclusion criteria included the following: (i) studies not providing data for OR calculation between dyslipidemia, its parameters (TG, HDL, LDL, TC) and urolithiasis; (ii) review or meta-analysis studies; (iii) comments, editorials, case reports, letters and meeting/congress abstracts; and (iv) animal experiments.
The term, dyslipidemia, was used as an increased TG and decreased HDL level in individual studies. Given that TG and HDL are the two main components of Mets, we tried to com-bine them separately instead of only calculating the association of dyslipidemia and urolithiasis. However, some studies just established the relationship of dyslipidemia and urolithiasis, and there was no detailed information about TG and HDL level, so we also made a calculation for the association of dys-lipidemia (any disturbances in lipid profile) and urolithiasis.
Data included for each study were as follows: the first author’s last name, year of publication, study design, country of study, population, time period, cohort size including males and females, age range, dyslipidemia traits, stone composition,
adjustments for the calculation of dyslipidemia and urolithiasis association, and quality assessment of the studies.
Quality assessment
The included studies were assessed by the NOS.17 The NOS is judged on three broad subscales using a star to identify high-quality choices: the selection of the study groups con-tains four items; the comparability of the groups comprises two elements, and the ascertainment of the exposure; or out-come of interest for observational studies includes three items. A score of≤5 was regarded as low quality.
Statistical analysis
We carried out meta-analysis pooling of the relevant studies that met the inclusion criteria. Statistical heterogeneity among trials was assessed by using Cochran’s Q and I2
statistic, and a P-value <0.05 or I2 value >50% was considered to be heterogeneous.18 As individual studies were gathered from published literature and they did not share common effect size, we used a random effects model. We carried out sensi-tivity analysis to assess the robustness of the results in the present meta-analysis. The purpose of the sensitivity analysis was to evaluate the effect of a single study on the overall pooled estimates. The sensitivity analysis was carried out by omitting one study at a time, generating the pooled estimates and comparing with the original estimates. We assessed the presence of publication bias using funnel plots, which display the relationship between study size and effect size. To quan-tify the relationship, Begg and Mazumdar rank correlation and Egger’s regression intercept tests were used.19,20 We fur-ther evaluated publication bias using Orwin’s fail-safe N method. All statistical comparisons were two-sided, and a P-value<0.05 was considered statistically significant. All analy-ses were carried out using comprehensive meta-analysis Version.3 (Biostat, Englewood, NJ, USA).
Results
Literature search and study characteristics
We initially identified 719 studies, either in full publications or abstract forms, using the methodology and the search terms described above. Of the studies, 91 were reviews, 242 were animal model studies, 61 were child population studies and 230 had outcomes not relevant to the subject. A total of 95 publications were retrieved for further evaluation. Of these, 60 case reports, and 13 letters and editorials were excluded. Of the remaining 22 studies, 11 were eliminated because they did not report the mean value with standard errors or OR with 95% CIs or provide sufficient data to calculate them. Finally, 11 studies were included in the present meta-analysis. The details of the literature search are shown in Figure 1.
The 11 selected studies contained 282 479 participants (ranging from 694 to 116 536).21–31 They were published between 2008 and 2017. Of these, seven were cross-sectional studies, two were retrospective cohort studies, one was a prospective cohort study and one was a case–control study. As multivariate analysis of potential confounders is more
informative, giving more profound and robust insight for the association, instead of crude values, adjusted OR (for cross-sectional and case report studies) or HR (for cohort studies) were pooled in the analysis. The regional distribution of the studies included were as follows: nine from Asia, one from Europe and one from the USA. Eight studies reported the association between TG and urolithiasis,21–28 whereas nine studies reported the relationship between HDL and urolithia-sis.21–29Three studies showed the association of any dyslipi-demia traits (not separately reported TG, HDL, TC, LDL) with urolithiasis.29–31The detailed characteristics of included studies are presented in Table 1.
The cut-off values of dyslipidemia traits and diagnostic tools for urolithiasis detection are shown in Table 2. All indi-vidual studies used 150 mg/dL for a high TG cut-off value,
except for the study by Masterson et al., in which no detailed information about TG was available.29 Three studies only included men, thus only HDL for men <40 mg/dL was reported as a cut-off value in those studies.25,26,31Kohjimoto et al. reported 40 mg/dL of HDL as the cut-off value for both sexes.30Masterson et al. used cut-off values of <45 mg/
dL for men and <60 mg/dL for women.29 The remaining studies were the same using HDL values of <40 mg/dL for men and<50 mg/dL for women.
The most common diagnostic tool used for the detection of urinary stone was USG. Jeong et al.22 carried out USG plus
CT, whereas Kang et al.24used USG plus abdominal radiog-raphy. Ando et al. used detailed questionnaire in addition to USG.31 Masterson et al. evaluated medical records screening relevant codes for urolithiasis.29Kohjimoto et al. only used a
Records identified through database searching (n = 719) Identification Scr eening Eligibility Included
Additional records identified through other sources
(n = 0)
Studies retrieved for further evaluation
(n = 95)
Potential studies fulfilling the inclusion criteria
(n = 22)
Full-text articles excluded due to
lack of sufficient data (n = 11) Excluded after reading the full-texts: • Case report (n = 60)
• Letter and editorials (n = 13)
Records excluded by reading title and abstracts • Children population (n = 61) • Animal studies (n = 242)
• Not report the relationship between • Reviews (n = 91)
dyslipidemia and urolithiasis (n = 230)
Studies included in quantitative synthesis (meta-analysis)
(n = 11)
• Cross-sectional studies (n = 7) • Case-control studies (n = 1) • Cohort studies (n = 3)
Table 1 Relevant studies on the association of blood lipid profile and urolithiasis Authors (publication year) Study design Country Population Time pe riod To tal cohort size (m ale/female) Age range (years) or mea n age SD Participants with NL/non-NL Dyslipid emia traits OR Stone composition Adjusted covariates Quality Rendina et al. (2011 ) 21 Cross-sec tional Italy Caucasia n inpatients 2004 –2005 2132 (1038/1094) 63.8 15.8 (62.9 15.7/ 64.7 15.9) 220/1912 TG HDL 1.4 (0.86 –2.25 ) 1.3 (0.75 –2.15) NA Age, waist circumference hypertension ,
abnormal glucose metabolism
6 Jeong et al. (2011) 22 Cross-sec tional Korea Asian Medical Cente r 2006 34 895 (20 790/14 105) ≥ 20 839/34 056 TG HDL 1.07 (0.86 –1.35 ) 0.88 (0.70 –1.10 ) NA Age, sex 7 Kim et al. (2013) 23 Cross-sec tional Korea Health promotion center 2010 in Seoul, Korea 2010 116 536 (67 262/49 274) 45.3 9 8.74 (45.56 8.61 / 45.05 9.0) 7107/109 429 TG HDL 1.03 (0.97 –1.10 ) 0.97 (0.89 –1.06 ) NA Age, creatinine, uric acid level, past medical history of NL 8 Kang et al. (2014 ) 24 Case –control Korea Cases – medic al records Controls – health promotion center of the hospita l 2005 –2011 2620 655/1965 TG HDL TC, LDL 1.90 (1.58 –2.28 ) 1.88 (1.54 –2.30 ) NA Obesity, diabetes, hypertension 7 Ca se (405/250) 46.8 13.2 Contr ol (122 6/739) 46.8 11.5 Lee et al. (2016) 25 Cross-sec tional Taiwan Taiwa nese men undergoing free health NA 694 43 –83 (55.6 4.6) 85/609 TG HDL 1.53 (0.94 –2.50 ) 1.39 (0.85 –2.25 ) NA Age, testosterone level 7 Chang et al. (2011) 26 Prosp ective cohort Korea Worker s participated in a comprehen sive health examination 2002 –2009 3872 41.6 7.2 118/3754 TG HDL 1.91 (1.31 –2.78 ) 1.0 (0.69 –1.46) NA Age, GFR, uric ac id level, incidental hypertensio n, diabetes mellitus 6 Jung et al. (2011) 27 Retrospec tive cohort Korea Korea n participants from a health promotion center 1995 –2009 40 687 (22 540/18 147) 44.9 11.5 (44.5 11.3/ 45.5 11.8) 609/40 078 TG HDL 1.07 (0.93 –1.22 ) 1.12 (1.02 –1.22 ) NA Age, GFR, serum u ric acid, creatinine level 6 Liu et al. (2017 ) 28 Cross-sec tional Taiwan Health Examination Cente r o f Changh ua Christian Hospital 2010 3793 (2118/1675) 46.4 9 11.91 639/3154 TG HDL LDL 1.08 (0.88 –1.33 ) 1.12 (0.94 –1.35 ) NA Age, sex, serum uric acid, creatinine level 7 Masterson et al. (2015) 29 Retrospec tive cohort USA Active duty military personnel, retirees, and th eir de pendents 2000 –2012 52 184 (29 467/22 717) 31.0 15.3 702/51 482 HDL Dyslipid emia 1.30 (1.02 –1.64 ) 1.20 (1.0 –1.44) NA NA 7 Kohjimoto et al. (2013) 30 Cross-sec tional Japan 6th Natio nwide Survey on Urolithia sis 2005 11 555 (8534/3021) 52.5 14.4 6603/4952 Dyslipid emia 1.36 (1.22 –1.52 ) N A Age, sex, obesity , diabetes, hypertension 7 Ando et al. (2013) 31 Cross-sec tional Japan Japanese men undergoing comprehen sive health medical examina tion 1995 –2001 13 418 404/13 014 Dyslipid emia 1.28 (1.05 –1.56) NA Age, smoking, alcohol consumptio n, exercise 6 Contr ol (11 783) 48.5 8.8 Curr ent SF (404) 50.1 8.4 Previo us SF (123 1) 49.7 8.4
detailed questionnaire for kidney stone detection.30 The mean quality of cross-sectional studies was 6.8; cohort studies were 6.3 and case–control study was 7. The details for the quality assessment of the manuscripts including total score as well as subscores of selection, comparability and outcome are shown in Table 3.
High TG level and urolithiasis
The pooled analysis of OR of eight individual studies showed that patients with high TG levels had a higher overall adjusted estimated risk of urolithiasis (1.287 [1.073–1.544],
Q= 49.731; P-value for heterogeneity <0.0001;
I2 = 85.924%).21–28Figure 2 shows the details of the individ-ual studies and pooled analysis including OR and CI calcula-tions with their relative weights.
Low HDL level and urolithiasis
The pooled analysis of OR of nine individual studies showed that patients with low HDL levels had a higher overall adjusted estimated risk of urolithiasis (1.173 [1.014–1.358],
Q= 42.945; P-value for heterogeneity <0.0001;
I2 = 81.371%).21–29Figure 3 shows the details of the individ-ual studies and pooled analysis including OR and CI calcula-tions with their relative weights.
Any component of dyslipidemia and urolithiasis
The pooled analysis of OR of three individual studies showed that patients with any dyslipidemia traits levels had a higher
overall adjusted estimated risk of urolithiasis (1.309 [1.202–
1.425], Q= 1.387; P-value for heterogeneity = 0.5;
I2 = 0.000%).29–31Figure 4 shows the details of the individ-ual studies and pooled analysis including OR and CI calcula-tions with their relative weights.
Subgroup analysis
Subgroup analysis was carried out according to the region and study type for the association between HDL and TG urolithiasis separately. For the association between high TG level and urolithiasis, the pooled OR for cross-sectional
Table 2 Cut-off values for dyslipidemia traits and methods for urolithiasis diagnosis used in included studies
Authors (publication year) Dyslipidemia traits Diagnostic tools for urolithiasis
Rendinaet al. (2011)21 TG>1.7 mmol/L
HDL for men<1.03 mmol/L
HDL for men<1.3 mmol/L
Ultrasonography
Jeonget al. (2011)22 TG≥150 mg/dL
HDL for men<40 mg/dL
HDL for women<50 mg/dL
Ultrasonography and CT
Kimet al. (2013)23 TG≥150 mg/dL or treatment for high TG
HDL for men<40 mg/dL or treatment for low HDL
HDL for women<50 mg/dL or treatment for low HDL
Ultrasonography
Kanget al. (2014)24 TG≥150 mg/dL
HDL for men<40 mg/dL
HDL for women<50 mg/dL
Abdominal radiography and ultrasonography
Leeet al. (2016)25 TG≥150 mg/dL
HDL for men<40 mg/dL
Findings from medical records
Evidence of kidney stone from ultrasonography Operative history of stones removal from kidney
Changet al. (2011)26 TG≥150 mg/dL HDL for men<40 mg/dL Ultrasonography Junget al. (2011)27 TG≥150 mg/dL HDL for men<40 mg/dL HDL for women<50 mg/dL Ultrasonography Liuet al. (2017)28 TG≥150 mg/dL HDL for men<40 mg/dL HDL for women<50 mg/dL Ultrasonography
Mastersonet al. (2015)29 HDL for men<45 mg/dL
HDL for women<60 mg/dL
Findings from medical records using relevant codes
Kohjimotoet al. (2013)30 TG≥150 mg/dL
HDL<40 mg/dL
Detailed questionnaire including stone location, size and composition
Andoet al. (2013)31 TG≥150 mg/dL
HDL for men<40 mg/dL
Ultrasonography
Questionnaire including medical history of kidney stone
Table 3 NOS results for quality evaluation
Authors (publication year) Selection Comparability Outcome Total
Rendinaet al. (2011)21 *** * ** 6* Jeonget al. (2011)22 **** ** * 7* Kimet al. (2013)23 **** ** ** 8* Kanget al. (2014)24 **** ** * 7 Leeet al. (2016)25 **** * ** 7* Changet al. (2011)26 *** ** * 6* Junget al. (2011)27 *** * ** 6* Liuet al. (2017)28 **** ** * 7* Mastersonet al. (2015)29 **** ** * 7* Kohjimotoet al. (2013)30 **** * ** 7* Andoet al. (2013)31 *** ** * 6*
studies was 1.05 (0.98–1.12); P = 0.13, whereas for cohorts the pooled OR was 1.39 (0.79–2.44); P = 0.25. According to the region, the pooled OR was 1.27 (1.05–1.54); P = 0.01 for Asian populations and 1.40 (0.86–2.26); P = 0.17 for European populations.
For the association between low HDL level and urolithia-sis, the pooled OR for cross-sectional studies was 1.01 (0.91– 1.13); P= 0.75, whereas for cohorts the pooled OR was 1.13 (1.04–1.23); P = 0.002. According to the region, the pooled OR was 1.15 (0.97–1.35); P = 0.1 for Asian, 1.30 (0.76– 2.10); P= 0.32 for European and 1.30 (1.02–1.64); P = 0.02 for American populations.
The detailed subgroup analysis is shown in Table 4.
Sensitivity analysis
Sensitivity analysis for the association between high TG, low HDL level and estimated risk of urolithiasis was carried out to evaluate the robustness of our meta-analysis. Each study was excluded in turn to recalculate the pooled OR of the remaining studies. For the association between high TG and
estimated risk of urolithiasis, no significant change was obtained in overall pooled ORs, which ranged from 1.14 (1.01–1.28) to 1.35 (1.08–1.69) after excluding any study. Similar to the cumulative analysis, evident heterogeneity was observed, and no single study dominated the pooled ORs and heterogeneity.
For the association between low HDL and estimated risk of urolithiasis, a significant change was observed in overall pooled ORs, which ranged from 1.06 (0.97–1.16) to 1.21 (1.02–1.44) after excluding the studies of Rendina et al.,21 Kang et al.,24 Lee et al.,25 Masterson et al.,29Jung et al.27and Liu et al.28 one by one. For the association between any dyslipidemia traits and estimated risk of urolithiasis, no significant change was achieved in overall pooled ORs, which ranged from 1.23 (1.08– 1.41) to 1.34 (1.21–1.47) after excluding any study. The details are shown in Table 5.
Publication bias
The publication bias was evaluated for TG–urolithiasis and HDL–urolithiasis associations separately.
Model Study name Statistics for each study Odds ratio and 95% CI Weight (fixed) Relative weight Relative weight Relative weight Relative weight 1.09 100 10 1 0.1 0.01 0.170 1.371 0.510 0.921 6.885 1.704 0.798 1.000 3.381 4.078 2.720 1.544 1.073 1.056 1.313 0.935 0.938 0.884 1.583 0.967 0.847 0.866 1.400 1.060 1.030 1.901 1.530 1.088 1.072 1.912 1.110 1.287 Rendina et al.(21) Jeong et al.(22) Kim et al.(23) Kang et al.(24) Liu et al.(28) Jung et al.(27) Chang et al.(26) Fixed Random Meta Analysis Lee et al.(25) 1.168 2.784 1.229 1.339 2.495 2.282 1.097 1.326 2.264 Upper limit Lower limit Odds
ratio Z-value P-value 0.610 0.357 0.000 0.088 0.425 0.317 0.001 0.000 0.007 Favors A Favors B 7.79 13.61 16.80 14.61 7.65 14.02 15.63 9.89 5.05 63.95 7.56 1.06 5.89 13.62 1.79
Fig. 2 Meta-analysis plot for the association between high serum TG level and urolithiasis. ORs and 95% CIs of individual studies and of pooled data for the
asso-ciation between high TG level and urolithiasis in all participants (Q = 49.731; P-value for heterogeneity <0.0001; I2= 85.924%).
Model Study name
Rendina et al.(21) Jeong et al.(22) Kim et al.(23) Liu et al.(28) Jung et al.(27) Chang et al.(26) Masterson et al.(29) Kang et al.(24) Lee et al.(25) Fixed Random Odds ratio
Statistics for each study Lower
limit Upper
limitZ-valueP-value
Odds ratio and 95% CI Weight (fixed) Relative weight Relative weight Relative weight Relative weight 100 10 1 0.1 0.01 Favors A Favors B 0.99 5.25 12.46 15.59 11.59 12.58 5.83 13.18 15.52 8.00 6.53 35.95 5.00 6.78 1.16 8.41 33.17 1.99 Meta Analysis 1.300 1.300 1.390 1.128 1.124 1.007 1.098 1.173 1.886 0.900 0.768 0.977 0.733 0.329 0.314 0.495 0.028 0.000 0.185 0.191 0.012 0.032 0.971 0.889 1.028 1.105 1.326 1.307 –1.007 –0.683 1.059 1.643 1.351 1.231 1.460 1.157 1.358 2.201 2.194 2.518 6.180 2.306 2.262 1.542 0.854 0.942 1.026 1.042 1.014 2.144 0.695 0.037 3.507 0.970 0.000
Fig. 3 Meta-analysis plot for the association between low serum HDL level and urolithiasis. ORs and 95% CIs of individual studies and of pooled data for the
TG–urolithiasis
The funnel plot did not show apparent asymmetry (Fig. 5). The Begg and Mazumdar rank correlation test (P= 0.13) and Egger’s regression intercept test (P = 0.10) also showed no significant publication bias was available. Using 1.01 as a cri-terion for trivial OR, according to Orwin’s fail-safe N method calculation, 77 studies are required to bring the OR value to <1.01.
HDL–urolithiasis
The funnel plot did not show obvious asymmetry (Fig. 6). The Begg and Mazumdar rank correlation test (P= 0.53)
and Egger’s regression intercept test (P = 0.32) also showed no significant publication bias. Using 1.01 as a criterion for trivial OR, according to Orwin’s fail-safe N method calcula-tion, 73 studies are required to bring the OR value to <1.01.
Discussion
We carried out the present meta-analysis by pooling individ-ual studies to reach a more reliable and robust conclusion on the association between blood lipid profile and urolithiasis in which conflicting results are available in the literature. The term dyslipidemia mainly includes disturbances in TC, LDL, HDL and TG level. We tried to identify the association of urolithiasis with dyslipidemia traits separately as much as possible. We calculated TG–urolithiasis and HDL–urolithiasis
Model Study name Statistics for each study
Fixed Random Upper limit Lower limit Odds
ratio Z-value P-value
Meta Analysis
Masterson et al.(29) Kohjimoto et al.(30)
Ando et al.(31)
Odds ratio and 95% CI
100 10 1 0.1 0.01 Favors A Favors B Weight (fixed) Relative weight Relative weight Relative weight Relative weight 59.81 59.81 21.74 21.74 18.45 18.45 0.000 0.050 0.015 0.000 0.000 1.360 1.309 1.309 1.200 1.202 1.202 1.050 1.000 1.218 1.518 1.440 1.960 5.482 2.444 6.204 6.204 1.425 1.425 1.560 1.280
Fig. 4 Meta-analysis plot for the association between dyslipidemia (any traits) and urolithiasis. ORs and 95% CIs of individual studies and of pooled data for the
association between dyslipidemia and urolithiasis in all participants (Q = 1.387; P-value for heterogeneity = 0.5; I2= 0.000%).
Table 4 Subgroup analysis of the association between lipid parameters and urolithiasis Group No. study OR (95% CI) Heterogeneity P-value for difference P-value I2(%) TG–urolithiasis 8 1.28 (1.07–1.54) <0.001 85.92 0.007 Study design Cross-sectional 5 1.05 (0.98–1.12) 0.39 2.84 0.13 Cohort 2 1.39 (0.79–2.44) 0.005 87.59 0.25 Case–control 1 1.90 (1.58–2.28) <0.001 Geographic area Asia 7 1.27 (1.05–1.54) <0.001 87.25 0.01 Europe 1 1.40 (0.86–2.26) 0.17 HDL–urolithiasis 9 1.17 (1.01–1.35) <0.001 81.37 0.032 Study design Cross-sectional 5 1.01 (0.91–1.13) 0.23 28.49 0.755 Case–control 1 1.88 (1.54–2.30) <0.001 Cohort 3 1.13 (1.04–1.23) 0.54 0.0 0.002 Geographic area Asia 7 1.15 (0.97–1.35) <0.001 85.15 0.10 Europe 1 1.30 (0.76–2.20) 0.329 North America 1 1.30 (1.02–1.64) 0.02 Dyslipidemia– urolithiasis 3 1.30 (1.20–1.40) 0.50 0.0 <0.001 Study design Cross-sectional 2 1.34 (1.21–1.47) 0.60 0.0 <0.001 Cohort 1 1.20 (1.00–1.44) <0.001 Geographic area Asia 2 1.34 (1.21–1.47) 0.60 0.0 <0.001 North America 1 1.20 (1.00–1.44) <0.001
Table 5 Sensitivity analysis after each study was excluded in turn
Study omitted OR (95% CI) for remainders P-value for difference Heterogeneity P-value I2(%) TG–urolithiasis Rendinaet al. (2011)21 1.27 (1.05–1.54) 0.012 <0.0001 87.71 Jeonget al. (2011)22 1.33 (1.08–1.64) 0.007 <0.0001 87.89 Kimet al. (2013)23 1.35 (1.08–1.69) 0.008 0.001 82.63 Kanget al. (2014)24 1.14 (1.01–1.28) 0.027 0.032 56.50 Leeet al. (2016)25 1.26 (1.04–1.53) 0.014 <0.0001 87.51 Changet al. (2011)26 1.23 (1.02–1.47) 0.026 <0.0001 85.55 Junget al. (2011)27 1.34 (1.06–1.69) 0.013 <0.0001 87.86 Liuet al. (2017)28 1.32 (1.07–1.64) 0.008 <0.0001 87.92 HDL–urolithiasis Rendinaet al. (2011)21 1.16 (0.99–1.35) 0.053 <0.0001 83.21 Jeonget al. (2011)22 1.21 (1.03–1.42) 0.016 <0.0001 81.75 Kimet al. (2013)23 1.21 (1.02–1.44) 0.027 <0.0001 77.07 Kanget al. (2014)24 1.06 (0.97–1.16) 0.163 0.103 41.27 Leeet al. (2016)25 1.15 (0.99–1.35) 0.061 <0.0001 82.99 Mastersonet al. (2015)29 1.15 (0.98–1.35) 0.069 <0.0001 82.86 Changet al. (2011)26 1.18 (1.01–1.39) 0.033 <0.0001 83.29 Junget al. (2011)27 1.18 (0.97–1.44) 0.089 <0.0001 83.17 Liuet al. (2017)28 1.18 (0.99–1.39) 0.056 <0.0001 83.33 Dyslipidemia–urolithiasis Mastersonet al. (2015)29 1.34 (1.21–1.47) <0.0001 0.60 0.0 Kohjimotoet al. (2013)30 1.23 (1.08–1.41) 0.002 0.63 0.0 Andoet al. (2013)31 1.30 (1.16–1.46) <0.0001 0.24 24.68
separately. Three studies presented their relationship using general term dyslipidemia, and we also estimated the dyslipi-demia urolithiasis association.29–31
Given the high prevalence of detection and the recurrence rate of urolithiasis, besides the conventional preventive meth-ods, new therapeutic and preventive strategies should be uti-lized as much as possible in light of current scientific evidence. According to our overall pooled analysis, a separate association of high TG and low HDL levels with urolithiasis was found to be significant. Additionally, a strong association was found between dyslipidemia (disturbances in any serum lipid traits) and the estimated risk of urolithiasis. However, we should also take into account the subgroup analysis of TG, HDL and urolithiasis association. The subgroup analysis showed conflicting results. These results should be interpreted with caution. Dyslipidemia is a term used for patients with either enhanced TC, LDL, TG level or decreased HDL levels. Thus, the prevalence of dyslipidemia is likely to be higher compared with that of one component of dyslipidemia. This could be an explanation of the outcomes of our analysis. Additionally, given dyslipidemia is an integral part of Mets, these results suggest that stone formers should be evaluated regarding a full lipid panel.
There is a bidirectional association between urolithiasis and most cardiovascular risk factors, including hypertension
and diabetes mellitus.32 Dyslipidemia is closely related to atherosclerosis, which is the main risk factor for cardiovascu-lar diseases. Systemic diseases and urolithiasis share simicardiovascu-lar metabolic responses and common pathophysiological mecha-nisms. Dyslipidemic patients are more likely to be overweight or obese. Obese individuals are found to have lower urinary pH, which is crucial for calcium oxalate and uric acid stone crystallization. Insulin resistance, the basic entity of Mets, might be of great importance as an explanatory factor for the associations between diabetes mellitus, obesity, dyslipidemia and renal stone disease. Previous studies suggested that urine pH and ammonium significantly decrease with an increasing number of Mets components.33,34 In an animal experiment and cell culture model study, Bobulescu et al. suggested that these responses might be associated with the lipid accumula-tion within the kidney, highlighting the relevance of lipid metabolism involved in the mechanisms of kidney stone for-mation.35
Lifestyle habits including diet, exercise and smoking are partially well-defined, but still not a fully uncovered issue for the risk of stone formation. Naya et al. assessed various diet-ary fatty acids, animal fat and animal protein in 58 idiopathic stone formers in their fourth decade, and reported that the nutritional content of arachidonic acid was positively corre-lated with urinary oxalate excretion.36 A correct bodyweight,
Funnel plot of standard error by log odds ratio 0.0 0.1 0.2 0.3 0.4 –2.0 –1.5 –1.0 –0.5 0.0 0.5 1.0 1.5 2.0 Log odds ratio
Standard err
or
Fig. 5 Funnel plot of publication bias for the association between TG and urolithiasis.
Funnel plot of standard error by log odds ratio
Log odds ratio
Standard err or 0.0 0.1 0.2 0.3 0.4 –2.0 –1.5 –1.0 –0.5 0.0 0.5 1.0 1.5 2.0
Fig. 6 Funnel plot of publication bias for the association between HDL and urolithiasis.
regular exercise and a reduction in stressful life events are useful actions in addition to a proper diet including a high intake of fluids, fruits and vegetables, low consumption of salt and protein, and limited carbohydrate.37 Exercise improves insulin sensitivity, the cornerstone of the association between dyslipidemia and urolithiasis, and is expected to improve Mets components.38,39
Renal tubular cell injury triggers crystal deposition, and lipid metabolites could be involved in the molecular mecha-nism of calcium oxalate crystallization in particular. Tsujihata et al. showed, in an experimental study, that atorvastatin has an inhibitory effect on renal tubular injury and oxidative stress caused by oxalate crystals, and concluded atorvastatin could help to prevent and treat crystal formation.40 This ani-mal model study was supported by a clinical perspective, as Sur et al. showed that statin medications had a protective effect against stone formation after adjusting for age, sex and comorbidities.41 The results of the present analysis also pro-vide some support for the use of antilipidemic drugs for urolithiasis, but we cannot yet suggest the use of statins as a part of preventive strategies. Additionally, even if used, the duration and dosage of these drugs for urolithiasis are under debate, and future well-designed prospective studies are nec-essary in this regard.
Dyslipidemia status and related conditions are mostly linked to chronic inflammation and oxidative stress.42,43
Davalos et al. examined the effect of oxidative stress on cal-cium oxalate stone and the preventive effect of anti-oxidative agents in renal tubular epithelial LLC-PK1 cells culture.44 They concluded that oxidative stress appears to be a primary cytotoxic action of calcium oxalate monohydrate that can damage or kill renal cells, which could also lead to stone for-mation through the undefined cellular, physiological pro-cesses. Nephrolithiasis and accompanying comorbidities might trigger the oxidative stress cascade, which is the lead-ing factor in cell injury. Ultimately, high calcium and phos-phate or oxalate, and low citrate or magnesium in the urine might lead to crystallization in the collecting ducts of the kid-ney, which is oxidatively stressed and injured.1These
mecha-nisms are also suggestive of the potential links between lipid metabolism disorders and kidney stone formation.
The association of various types of stone formation with altered lipid metabolism is also an important topic. Scientific evidence is relatively sparse in this regard. Inci et al. carried out a case–control study comparing the lipid profiles of 49 stone formers and 50 randomized age and sex-matched con-trols.45 They showed that BMI, TC and TG levels were sig-nificantly higher in stone formers compared with the control group, and this association of BMI and TC with stone forma-tion was more prominent in those with uric acid and calcium oxalate dihydrate calculi than in calcium oxalate monohydrate calculi. Torricelli et al. also evaluated the possible link between dyslipidemia and 24-h urine analysis and the stone composition comprising data including 2923 participants, of whom 835 had stone composition available.46 They
con-cluded low HDL and high TGs are associated with lower uri-nary pH, and uric acid stones are more common in patients with increased TC and TGs. Furthermore, two retrospective cross-sectional studies showed that the great majority of
stone-forming patients with Mets produce calcium oxalate stones.47,48 They also showed that the percentage of uric acid stone formation was correlated with the accumulation of Mets components. We were unable to test these results in the present meta-analysis, as the individual studies lacked the stone composition and detailed 24-h urinalysis sufficient to be pooled. However, given that the basic com-ponents of the stones are calcium oxalate and uric acid, we might postulate that altered lipid components could be asso-ciated with these types of urinary stones. It is that clear future studies should also focus on the possible links between lipid metabolism profiling and various types of stone formations.
Given the studies included in the meta-analysis were all observational, we implemented the meta-analytic approaches strictly in allfields of the analysis process, including eligibil-ity of studies, pooling and outcomes evaluation. Although the overall pooled report seems to show consistent results regard-ing the association between lipid parameters and urolithiasis, the subgroup analysis of TG– and HDL–urolithiasis associa-tion had conflicting findings. There are influential messages of this analysis. The stone formers should be evaluated regarding the entire content of the dyslipidemic panel (TG, HDL, LDL, TC). These results also suggest that a holistic approach is essential for stone formers comprising all individ-ual components of Mets, including dyslipidemia, hyperten-sion, impaired glucose metabolism and central obesity. The large number of participants (n = 282 479) included enabled the improvement of the precision of risk estimates, and allowed us to make conclusions based on this meta-analysis more precise and robust. In addition, we pooled multivari-able-adjusted risk estimates to minimize the confounding fac-tors.
There are several limitations to this analysis. The observa-tional studies reflect only individual conditions, and cannot permit us to form a cause and effect relationship, which necessitates further prospective studies. Evident heterogeneity was found in all analyses. The different calendar period and baseline characteristics of each study population, differences for the cut-off value of dyslipidemia traits and kidney stone detection tools, and the type and extent of statistical adjust-ment for confounders in the individual studies might be the potential relevant sources of heterogeneity. The number of studies might seem insufficient, but we believe that rather than the number of studies included, a robust statistical analy-sis focusing not only on summary effect size (random or fixed), but also evaluating different aspects of the report, including heterogeneity, effect size diversion, publication bias calculation and weight distribution of the studies, could pro-vide us more accurate meta-analysis evaluation. Furthermore, we were unable to uncover the possible association of dys-lipidemia with urolithiasis according to sex stratification because of insufficient knowledge in individual studies. We were also unable to make a subgroup analysis for stone type classification, as the data from the individual studies were inadequate. Finally, we could not calculate the association of TC and LDL cholesterol with urolithiasis because of insuffi-cient data in the studies, most of which were regarding the association between Mets and urolithiasis.
Consequently, the present meta-analysis showed that patients with higher TG and lower HDL had an increased estimated risk of urolithiasis. The TG–urolithiasis association was found to be more coherent and consistent compared with the HDL–urolithiasis association. These results should be interpreted with caution for clinical relevance. Well-designed clinical prospective studies examining the alleviating effect of antilipidemic drugs on the urolithiasis disease process could perhaps give more insight into understanding this association. Nevertheless, these data provide enough preliminary knowl-edge for urologists to handle urolithiasis as the manifestation of a systemic disorder implying that patients with urolithiasis should be encouraged with lifestyle modifications due to the close relationship between urolithiasis and dyslipidemia.
Acknowledgments
No research support or funding was received in connection with this study. The authors have no significant affiliation or involvement, either direct or indirect, with any organization or entity with a direct financial interest in the subject matter or materials discussed.
Conflict of interest
None declared.
References
1 Khan SR. Is oxidative stress, a link between nephrolithiasis and obesity, hypertension, diabetes, chronic kidney disease, metabolic syndrome? Urol. Res. 2012; 40: 95–112.
2 Scales CD Jr, Smith AC, Hanley JM, Saigal CS. Prevalence of kidney stones in the United States. Eur. Urol. 2012; 62: 160–5.
3 Pearle MS, Calhoun EA, Curhan GC. Urologic diseases in America project: urolithiasis. J. Urol. 2005; 173: 848–57.
4 Curhan GC, Willett WC, Rimm EB, Speizer FE, Stampfer MJ. Body size and risk of kidney stones. J. Am. Soc. Nephrol. 1998; 9: 1645–52. 5 Asplin JR. Obesity and urolithiasis. Adv. Chronic Kidney Dis. 2009; 16: 11–20. 6 Negri AL, Spivacow FR, Del Valle EE, Forrester M, Rosende G, Pinduli I. Role of overweight and obesity on the urinary excretion of promoters and inhibitors of stone formation in stone formers. Urol. Res. 2008; 36: 303–7. 7 Taylor EN, Stamper MJ, Curhan GC. Diabetes mellitus and the risk of
nephrolithiasis. Kidney Int. 2005; 68: 1230–5.
8 Liu LH, Kang R, He J, Zhao SK, Li FT, Zhao ZG. Diabetes mellitus and the risk of urolithiasis: a meta-analysis of observational studies. Urolithiasis 2015; 43: 293–301.
9 Rezaee ME, Ward CE, Pollock M, Shetty SD. Association between multiple chronic conditions and urolithiasis. Int. Urol. Nephrol. 2017; 49: 1361–7. 10 Obligado SH, Goldfarb DS. The association of nephrolithiasis with
hyperten-sion and obesity: a review. Am. J. Hypertens. 2008; 21: 257–64.
11 Wong Y, Cook P, Roderick P, Somani BK. Metabolic syndrome and kidney stone disease: a systematic review of literature. J. Endourol. 2016; 30: 246– 53.
12 Rendina D, De Filippo G, D’Elia L, Strazzullo P. Metabolic syndrome and nephrolithiasis: a systematic review and meta-analysis of the scientific evi-dence. J. Nephrol. 2014; 27: 371–6.
13 West B, Luke A, Durazo-Arvizu RA, Cao G, Shoham D, Kramer H. Meta-bolic syndrome and self-reported history of kidney stones: the National Health and Nutrition Examination Survey (NHANES III) 1988–1994. Am. J. Kidney Dis. 2008; 51: 741–7.
14 Khan SR. Hyperoxaluria-induced oxidative stress and antioxidants for renal protection. Urol. Res. 2005; 33: 349–57.
15 Schmiedl A, Schwille PO, Bonucci E, Erben RG, Grayczyk A, Sharma V. Nephrocalcinosis and hyperlipidemia in rats fed a cholesterol- and fat-rich
diet: association with hyperoxaluria, altered kidney and bone minerals, and renal tissue phospholipid-calcium interaction. Urol. Res. 2000; 28: 404–15. 16 Fujii Y, Okada A, Yasui T et al. Effect of adiponectin on kidney crystal
for-mation in metabolic syndrome model mice via inhibition of inflamfor-mation and apoptosis. PLoS ONE 2013; 22: e61343.
17 Stang A. Critical evaluation of the Newcastle–Ottowa Scale for the
assess-ment of the quality of nonrandomized studies in meta-analysis. Eur. J. Epi-demiol. 2010; 25: 603–5.
18 Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557–60.
19 Egger M, Davey SG, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 629–34.
20 Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994; 50: 1088–101.
21 Rendina D, Mossetti G, De Filippo G et al. Association between metabolic syndrome and nephrolithiasis in an inpatient population in southern Italy: role of gender, hypertension and abdominal obesity. Nephrol. Dial. Transplant. 2009; 24: 900–6.
22 Jeong IG, Kang T, Bang JK et al. Association between metabolic syndrome and the presence of kidney stones in a screened population. Am. J. Kidney Dis. 2011; 58: 383–8.
23 Kim YJ, Kim CH, Sung EJ, Kim SR, Shin HC, Jung WJ. Association of nephrolithiasis with metabolic syndrome and its components. Metabolism 2013; 62: 808–13.
24 Kang HW, Lee SK, Kim WT et al. Hypertriglyceridemia and low high-den-sity lipoprotein cholesterolemia are associated with increased hazard for urolithiasis. J. Endourol. 2014; 28: 1001–5.
25 Lee YC, Huang SP, Juan YS, Huang TY, Liu CC. Impact of metabolic syn-drome and its components on kidney stone in aging Taiwanese males. Aging Male 2016; 19: 197–201.
26 Chang IH, Lee YT, Lee DM et al. Metabolic syndrome, urine pH, and time-dependent risk of nephrolithiasis in Korean men without hypertension and diabetes. Urology 2011; 78: 753–8.
27 Jung HS, Chang IH, Kim KD et al. Possible relationship between metabolic syndrome traits and nephrolithiasis: incidence for 15 years according to gen-der. Korean J. Urol. 2011; 52: 548–53.
28 Liu YT, Yang PY, Yang YW, Sun HY, Lin IC. The association of nephrolithiasis with metabolic syndrome and its components: a cross-sectional analysis. Ther. Clin. Risk Manag. 2017; 13: 41–8.
29 Masterson JH, Woo JR, Chang DC et al. Dyslipidemia is associated with an increased risk of nephrolithiasis. Urolithiasis 2014; 43: 49–53.
30 Kohjimoto Y, Sasaki Y, Iguchi M, Matsumura N, Inagaki T, Hara I. Associa-tion of metabolic syndrome traits and severity of kidney stones: results from a nationwide survey on urolithiasis in Japan. Am. J. Kidney Dis. 2013; 61: 923–9. 31 Ando R, Nagaya T, Suzuki S et al. Kidney stone formation is positively associated with conventional risk factors for coronary heart disease in Japa-nese men. J. Urol. 2013; 189: 1340–6.
32 Lange JN, Mufarrij PW, Wood KD, Holmes RP, Assimos DG. The associa-tion of cardiovascular disease and metabolic syndrome with nephrolithiasis. Curr. Opin. Urol. 2012; 22: 154–9.
33 Maalouf NM, Cameron MA, Moe OW et al. Low urine pH: a novel feature of the metabolic syndrome. Clin. J. Am. Soc. Nephrol. 2007; 2: 883–8. 34 Sakhaee K, Maalouf NM. Metabolic syndrome and uric acid nephrolithiasis.
Semin. Nephrol. 2008; 28: 174–80.
35 Bobulescu IA, Dubree M, Zhang J et al. Effect of renal lipid accumulation
on proximal tubule Na+
/H+
exchange and ammonium secretion. Am. J. Phys-iol. Renal PhysPhys-iol. 2008; 294: 1315–22.
36 Naya Y, Ito H, Masai M, Yamaguchi K. Association of dietary fatty acids with urinary oxalate excretion in calcium oxalate stone-formers in their fourth decade. BJU Int. 2002; 89: 842–6.
37 Meschi T, Nouvenne A, Borghi L. Lifestyle recommendations to reduce the risk of kidney stones. Urol. Clin. North Am. 2011; 38: 313–20.
38 Loprinzi PD, Fitzgerald EM, Woekel E et al. Association of physical activity and sedentary behavior with biological markers among U.S. pregnant women. J. Womens Health 2013; 22: 953–8.
39 O’Donovan G, Hillsdon M, Ukoumunne OC. Objectively measured physical
activity, cardiorespiratory fitness and cardiometabolic risk factors in the
Health Survey for England. Prev. Med. 2013; 57: 201–5.
40 Tsujihata M, Momohara C, Yoshioka I, Tsujimura A, Nonomura N, Okuyama A. Atorvastatin inhibits renal crystal retention in a rat stone form-ing model. J. Urol. 2008; 180: 2212–7.
41 Sur RL, Masterson JH, Palazzi KL et al. Impact of statins on nephrolithiasis in hyperlipidemic patients: a 10-year review of an equal access health care system. Clin. Nephrol. 2013; 79: 351–5.
42 Ozbek E. Induction of oxidative stress in kidney. Int. J. Nephrol. 2012; 2012: 465897.
43 Tugcu V, Ozbek E, Aras B, Arisan S, Caskurlu T, Tasci AI. Manganese superoxide dismutase (Mn-SOD) gene polymorphisms in urolithiasis. Urol. Res. 2007; 35: 219–24.
44 Davalos M, Konno S, Eshghi M, Choudhury M. Oxidative renal cell injury induced by calcium oxalate crystal and renoprotection with antioxidants: a pos-sible role of oxidative stress in nephrolithiasis. J. Endourol. 2010; 24: 339–45.
45 Inci M, Demirtas A, Sarli B, Akinsal E, Baydilli N. Association between body mass index, lipid profiles, and types of urinary stones. Ren. Fail. 2012; 34: 1140–3.
46 Torricelli FC, De SK, Gebreselassie S, Li I, Sarkissian C, Monga M. Dyslipi-demia and kidney stone risk. J. Urol. 2013; 191: 667–72.
47 Kadlec AO, Greco K, Fridirici ZC, Hart ST, Vellos T, Turk TM. Metabolic syndrome and urinary stone composition: what factors matter most? Urology 2012; 80: 805–10.
48 Cho ST, Jung SI, Myung SC, Kim TH. Correlation of metabolic syndrome with urinary stone composition. Int. J. Urol. 2013; 20: 208–13.
