Abstract
Renal dysfunction causes dyslipidemia, and the progression of kidney disease is associated with treatment-requiring lipid disorders. However, there is little data on the association between the severity of dyslipidemia and various treatment modalities with renal function. A total of n = 214 prevalent patients from a lipidology and nephrology practice were investigated in an unicentric cross-sectional study and divided into 4 groups based on the severity and therapeutic regimen of an existing lipid metabolism disorder: a lipid apheresis-treated group (LA), a drug-treated group (MG), a control group (CG) that included patients with lipid disorder not needing any treatment, and a diabetes group that comprised all diabetes patients. We examined fat metabolism parameters, renal parameters and urinary protein excretion, and compared these between study groups. Lipid apheresis therapy leads to the effective lowering of low density lipoprotein-concentration and to a reduction in the presumably protective high density lipoprotein-concentration. We found that S-crea was significantly higher in the LA, MG, and diabetes group than in the CG, while electronic glomerular filtration rate (eGFR) was correspondingly lower. On the other hand, proteinuria was not elevated in the LA- and MG-groups compared to CG. Multiple regression analysis revealed lipoprotein (a) values as a predictor for eGFR decline and increased proteinuria. Elevated lipoprotein (a) levels are associated with decreased eGFR and worse proteinuria, highlighting that regular monitoring and effective treatment, to date especially lipid apheresis, are of utmost importance to mitigate detrimental effects on renal function.
Keywords: albuminuria, dyslipidemia, lipid apheresis, Lp(a), medical treatment, proteinuria, renal function
1. Introduction
Dyslipidemia is an established risk factor for cardiovascular diseases. Patients with chronic kidney disease also carry an increased risk of cardiovascular morbidity and mortality, which is even exacerbated in case of an additional lipid metabolism disorder. Existing albuminuria aggravates the cardiovascular risk profile further.[1–3] Of note: dyslipidemia triggers an additional deterioration in renal function (an effect called lipid nephrotoxicity that is defined as an accumulation of lipid metabolites); it has an impact on mesangial cell proliferation, podocyte injuries, and tubulointerstitial disease.[4]
The definition of dyslipidemia comprises an abnormal range of lipoproteins such as high density lipoprotein-concentration (HDL-C), low density lipoprotein-concentration (LDL-C), lipoprotein (a) (Lp(a)), total cholesterol (TC), and triglycerides. Dyslipidemia can cause endothelial dysfunction favoring the formation of atherosclerotic plaques.[5] LDL-C concentrations correlate with the amount of atherosclerotic plaques.[4,6]
HDL-C function remains controversial. Initially, it was believed that low HDL-C levels indicated a higher cardiovascular risk. However, several recent studies have described HDL-C’s role in cardiovascular events to be ambivalent.[7] As renal dysfunction progresses, the concentrations of TC and LDL-C tend to fall, HDL-C is also reduced, while TG increases.[8] In patients with an electronic glomerular filtration rate (eGFR) > 90 mL/min/1.73 m2, a high HDL-C concentration correlates with fewer cardiovascular events, but this correlation is not observed in patients with an eGFR < 90 mL/min/1.73 m2.[8]
Lp(a) is an independent cardiovascular risk factor and is strongly associated with mortality due to cardiovascular events.[9] Lp(a) should be measured at least once in a lifetime in every adult patient presenting a moderate or high cardiovascular risk profile to determine the risk of cardiovascular disease.[5]
The therapy of dyslipidemia depends on the cardiovascular risk constellation and includes life style modification and drug therapy, primarily initiated with statins. Ezetimibe can be used as an adjunct therapy, if the initial therapy with statins does not achieve the desired lipid levels or when higher doses of statins are not tolerated by the patient.[5] Intensified treatment approaches are proprotein convertase subtilisin/kexin type 9 therapy or lipid apheresis.[5,10,11] Lipid apheresis is an efficient therapy for patients with severe dyslipidemia carrying a high cardiovascular risk. It can be administered in patients with Lp(a) > 60 mg/dL and high cardiovascular risk, in homocygous familiar dyslipidemia, or in elevated LDL-C levels under maximum medical treatment.[5,10–12]
There is ample evidence that renal dysfunction leads to a progression of dyslipidemia, but there is a paucity of data on the impact of dyslipidemia on renal outcome, which motivated us to investigated this context. Therefore, we studied ambulant patients with dyslipidemia and divided them based on their dyslipidemia severity level and concomitantly treatment modality into 4 different groups: a lipid apheresis-treated group (LA), a drug-treated group (MG), a control group (CG) that included patients with lipid disorder not needing any treatment, and a diabetes group (DG) that comprised all diabetes patients. The rationale for the classification was based on different considerations: diabetes mellitus is known to have a significant impact on the lipid profile, therefore we allocated diabetes patients to a separate group. Additionally, elevated Lp(a) levels are not effectively lowered by conventional pharmacological therapies but can be treated by lipid apheresis. This led us to distinguish between patients treated by drugs and by lipid apheresis.
Overall, our study aims to investigate the influence of a severe dyslipidemia and its therapy on renal function.
2. Methods
2.1. Study design
This study was designed as an unicentric, cross-sectional study of prevalent patients with dyslipidemia (n = 214). The recruitment started in April 2016 at the North Rhine Dialysis and Lipid Center in conjunction with the department of nephrology of the University Hospital Essen, Germany. At this time-point, all patients undergoing lipid apheresis in the center (n = 72) were enrolled. Until January 2019 we included additional patients for the remaining study groups (described below), as they were less prevalent in the center. Clinical assessment included medical history, physical examination, as well as laboratory urine and blood tests. Blood tests included blood count, serum creatinine (s-crea), serum urea (s-urea), eGFR, HDL-C, LDL-C, Lp(a), TC, triglycerides (TAGs), C-reactive protein, alkaline phosphatase, total bilirubin, gamma glutamyl transferase (gGT), aspartate aminotransferase, and alanine aminotransferase. Urine analysis included urinary total protein (urinary tp), urinary total protein/crea ratio (urinary tp/crea), the urinary albumin/crea ratio, and urinary creatinine (u-crea). All parameters were assessed at the time of study inclusion, no follow-up assessments were conducted. TC, LDL, HDL, and TAG were measured by chromogenic photometry using reagents and automates by Beckmann Coulter, Brea. Lp(a) was measured by using a particle-enhanced immunoturbidimetry with the associated reagents by Diasys, Holzheim, Germany.
The study groups were defined as follows:
LA (n = 72): Patients with severe dyslipidemia undergoing lipid apheresis. Lipid apheresis was performed every 3 to 4 weeks. Blood values were taken immediately before apheresis.
MG (n = 54): Patients with dyslipidemia treated by lipid lowering drugs.
CG (n = 43): Patients with lipid disorder not needing treatment but regular checks of their lipid profile.
DG (n = 45): Patients with additional diabetes mellitus (DG) while presenting a lipid disorder.
Exclusion criteria were pregnancy, a prevalent chronic kidney disease with proven proteinuria (urinary tp > 3.5 g/24 h), drug therapy with proprotein convertase subtilisin/kexin type 9 antibodies and immunosuppressive therapy.
2.2. Statistical analysis
Data was analyzed by PASW Statistics (Version 22, SPSS, Chicago). Shapiro–Wilk test was used to test for normal distribution. After splitting the cohort into the study groups, all variables significantly deviated from a normal distribution. Data is presented as median with interquartile range. Numeric variables between 2 groups were compared via Mann–Whitney U test or by Kruskal–Wallis test for >2 groups, followed by Dunn post hoc test, corrected for multiple comparisons by Bonferroni. The Chi[2] served to compare categorical variables. Correlation analyses were subjected to Spearman correlation. Multiple linear regression was performed to test if age, body mass index (BMI), systolic blood pressure, hemoglobin A1C (HbA1c), and the lipid parameters could predict s-crea, eGFR, and urinary total protein/g creatinine ratio. Regression results were reported as standardized regression coefficient (β) and P value. All data were presented as median and interquartile range. P values < .05 were considered statistically significant with *P < .05, **P < .01, ***P < .001.
3. Results
Basic patient characteristics are summarized in Tables 1 and S1, Supplemental Digital Content, https://links.lww.com/MD/P798. The results are presented as median and compared in an intergroup analysis between the different study groups. The groups differed in age, as our CG patients were significantly younger than those in LA and MG. The distribution of sex was equal between the LA, MG, and DG with more male participants except for the CG with more female participants (70%).
Table 1.
Basic patient characteristics.
| n = 214 | LA (n = 72) | MG (n = 54) | CG (n = 43) | DG (n = 45) |
|---|---|---|---|---|
| Gender (%) | ||||
| F | 41.7 | 46.3 | 70 | 44.4 |
| M | 58.3 | 53.7 | 30 | 55.6 |
| Age | 57 (52–64.75) | 62 (51–68.25) | 49 (32–56) | 62 (52–71.5) |
| BMI (kg/m²) | 26 (23.7–29.55) | 25.95 (23.98–29.55) | 26.5 (21.8–30.1) | 30 (26.85–33.75) |
| Waist circumference | 98 (90–108) | 104 (94.75–115) | 102.5 (90.5–114.25) | 109 (102.5–121.25) |
| Systolic blood pressure (mm Hg) | 123.5 (117–135) | 132 (120–145) | 135 (120–150) | 130 (115–140) |
| HbA1c (%) | 5.5 (5.4–5.8) | 5.6 (5.4–5.9) | 5.35 (5.2–5.7) | 6.5 (6.1–7.65) |
| HDL-C (mg/dL) | 48 (42–59) | 49 (39–61) | 53 (43–68) | 42 (37.25–49.75) |
| LDL-C (mg/dL) | 105.5 (85–147.5) | 109.5 (96–135) | 149 (129–167) | 127 (93.25–149) |
| Lp(a) (mg/dL) | 72.5 (37.25–116) | 28 (10–118.25) | 10 (3–21.5) | 61 (8.5–116.25) |
| Total cholesterol (mg/dL) | 176 (150–223) | 176 (149–201) | 237 (212–267) | 190 (148.5–231.5) |
| Triglycerides (mg/dL) | 154 (117–220) | 130 (90–175) | 140 (105–281) | 198 (131–377.5) |
| S-Urea (mg/dL) | 38 (29–45) | 36 (27–45.5) | 29 (25–34) | 42 (29–57.5) |
| S-Creatinine (mg/dL) | 1.07 (0.93–1.16) | 0.97 (0.85–1.24) | 0.93 (0.82–1.05) | 0.99 (0.89–1.14) |
| eGFR (ml/min/1.73m2) | 68 (60–80.75) | 71 (58.5–84) | 80 (67–96) | 68 (53–84) |
| Na (mmol/L) | 139 (138–141.75) | 141 (139–142.2) | 141 (139–142) | 140 (138.5–141) |
| K (mmol/L) | 4.2 (3.92–4.6) | 4.5 (4.27–4.7) | 4.4 (4.2–4.62) | 4.3 (4.1–4.7) |
| Urinary total protein/creatinine (mg/g) | 63.5 (836–79) | 71.5 (51–91.75) | 83 (59–134.5) | 92 (58–161) |
| Urinary albumin/creatinine (mg/g) | 8 (5–16) | 12.5 (6.75–24.75) | 15.5 (9–47.25) | 16 (5–49) |
| Total bilirubin (mg/dL) | 0.5 (0.4–0.55) | 0.6 (0.5–0.8) | 0.4 (0.4–0.6) | 0.5 (0.4–0.8) |
| ALP (U/L) | 75.5 (61.25–87) | 78 (62–90) | 69 (60–91) | 80 (60.5–99.5) |
| gGT (U/L) | 29 (21.25–47.5) | 28.5 (19.75–42) | 21 (16–34) | 46 (29.5–93.5) |
| ALT (U/L) | 26.5 (23.25–32) | 26 (18–36.25) | 21 (17–34) | 32 (23.5–45.5) |
| AST (U/L) | 25 (20–33.75) | 27 (22–34.25) | 25 (21–29) | 27 (21–33) |
| TSH (U/L) | 1.5 (1.1–1.9) | 1.45 (1.0–2.18) | 1.72 (1.21–2.09) | 1.85 (1.1–2.4) |
Values indicate median (IQR).
ALP = alkaline phosphatase, ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, CG = control group, DG = diabetes group, gGT = gamma glutamyl transferase, HbA1c = hemoglobin A1C, HDL-C = high density lipoprotein-concentration, IQR = interquartile range, LA = lipid apheresis treated group, LDL-C = low density lipoprotein-concentration, MG = drug treated group.
All patients in the MG were treated with a statin: 53% received atorvastatin, 41% received simvastatin, 3% received rosuvastatin, and 3% received pravastatin. Ninety percent of the MG patients received an additional therapy with ezetimibe, while 15% received fibrates. The majority of patients (56%) in the DG suffered from insulin dependent diabetes mellitus (IDDM), while 44% were diagnosed with non-IDDM. A small subgroup of IDMM (n = 4 individuals) suffered from diabetes mellitus type 1. All other individuals form DG were diagnosed with diabetes mellitus type 2. Among oral antidiabetic drugs, metformin was prescribed to 47% of the patients, 18% received sitagliptin, 9% received SLGT2 inhibitors, 4% received glinides, and 2% received sulfonylurea derivates.
Concerning lipid profiles, we noted that HDL-C was significantly higher in CG, while LDL-C was also increased. Lipoprotein (a) was lowest in CG, whereas TC was increased in CG. On the other hand, TAGs were highest in DG (Fig. 1).
Figure 1.
Lipid profile of different study groups. Absolute values of HDL-cholesterol, LDL-cholesterol, total cholesterol, triglycerides, and Lp(a) were compared between study groups. Line indicates median. Comparisons by Kruskal–Wallis test with Dunn test for multiple comparisons. Correction for multiple comparisons by Bonferroni. *P < .05, **P < .01, ***P < .001. HDL = high density lipoprotein, LDL = low density lipoprotein, Lp(a) = lipoprotein (a).
Regarding renal parameters, s-crea and s-urea were lowest in CG, accompanied by increased eGFR (Fig. 2). Analysis of urinary parameters revealed higher urinary albumin/g creatinine and tp/g creatinine ratios in DG (Fig. 2).
Figure 2.
Kidney function of different study groups. Absolute values of s-creatinine, eGFR, s-urea, urinary albumin/creatinine, and urinary total protein/creatinine were compared between study groups. Line indicates median. Comparisons by Kruskal–Wallis test with Dunn test for multiple comparisons. Correction for multiple comparisons by Bonferroni. *P < .05, **P < .01, ***P < .001. eGFR = electronic glomerular filtration rate, s-urea = serum-urea.
Furthermore, we assessed the cardiovascular risk profile of the participants and compared them between the study groups. This revealed an increased BMI and waist circumference in DG compared to the other groups (Tables 1 and S1, Supplemental Digital Content, https://links.lww.com/MD/P798). The assessment of cardiovascular comorbidities displayed a comparable pattern in patients of the LA, MG, and CG with almost all patients suffering from hypertension and about 50% from coronary heart disease (Table 2). However, the prevalence of coronary artery disease was highest in the LA group (about 40%) compared to the MG (about 15%) and DG (about 24%). The control patients differed from the other groups as about half of them were diagnosed with hypertension and the prevalence of coronary heart disease and periphery artery disease was neglectable.
Table 2.
Cardiovascular comorbidities.
| %positive/group | LA | MG | CG | DG |
|---|---|---|---|---|
| Coronary heart disease | 45.8 | 51.8 | 0 | 46.7 |
| Periphery artery disease | 40.2 | 14.8 | 2.3 | 24.4 |
| Arterial hypertension | 100 | 88.9 | 48.8 | 97.8 |
CG = control group, LA = lipid apheresis treated group, MG = drug treated group.
Accordingly, an intergroup analysis of the intake of antihypertensive medication revealed the lowest intake rate in the CG (Table 3). Patients from all groups were preferably treated with angiotensin converting enzyme inhibitors/angiotensin II receptor type 1 receptor antagonists and/or beta blockers.
Table 3.
Anti-hypertensive medication.
| %positive/group | LA | MG | CG | DG |
|---|---|---|---|---|
| ACE inhibitors/AT1 receptor antagonists | 72 | 66.7 | 37.2 | 75.6 |
| Beta blocker | 82 | 59.3 | 23.3 | 77.8 |
| Ca antagonist | 23.6 | 24.1 | 14 | 42.2 |
| Diuretics | 47.2 | 24.1 | 16.3 | 60 |
ACE = angiotensin converting enzyme, AT1 = angiotensin II receptor type 1, CG = control group, DG = diabetes group, LA = lipid apheresis treated group, MG = drug treated group.
HbA1c was, as expected, highest in diabetes patients. In addition, we examined laboratory values of liver damage. Overall, individuals in DG had increased laboratory signs of liver damage, especially indicated by an elevated gGT. Notably, only 2 participants across all groups had diagnosed chronic inactive hepatitis B infection. Analysis of thyroid function by TSH determination showed no significant differences between the groups. Intake of L-thyroxin was comparable between DG (20%), LA (19%), and MG (15%), but lower in CG (9%).
Examination of gender differences revealed lower BMI-values (P = .008), waist circumference (P < .001), HbA1c (P < .001), s-urea (P < .001), and s-crea (P < .001) in women. In contrast, HDL-C and TC were significantly higher in women (P < .001).
To gain more insight into the impact of different types of diabetes mellitus on lipid and renal parameters in DG group, we performed a subgroup analysis between patients with IDDM and non-IDDM (Table S2, Supplemental Digital Content, https://links.lww.com/MD/P798). The comparison revealed a slight increase in creatinine in IDDM patients but no significant differences for all other compared markers. Overall, patients with IDDM tend to also have increased lipid values, but the small subgroups lack statistical power.
We also ran correlation analyses between urinary/renal parameters and risk factors for cardiovascular diseases (Table 4) and detected a significant positive correlation between age and the albumin/creatinine ratio (P = .015) as well as with renal retention parameters and a negative correlation with eGFR (P < .001). Systolic blood pressure correlated positively with the albumin/creatinine ratio (P = .006). Furthermore, we observed a positive correlation between both waist circumference and BMI with creatinine and urea.
Table 4.
Correlation coefficients between urinary/renal parameters and cardiovascular risk factors.
| Age | HbA1c | Systolic blood pressure | Waist circumference | BMI | ||
|---|---|---|---|---|---|---|
| Urinary total protein/g creatinine | r | .133 | .221** | .077 | .074 | .154 |
| Urinary albumin/creatinine | r | .166* | .183* | .192** | .076 | .041 |
| s-Urea | r | .477** | .283** | .071 | .289** | .305** |
| s-Creatinine | r | .337** | .251** | .113 | .281** | .234** |
| eGFR | r | −.594** | −.179** | −.072 | −.123 | −.129 |
Values indicate correlations; correlations were calculated by Spearman; r = correlation coefficient.
BMI = body mass index, eGFR = electronic glomerular filtration rate, HbA1c = hemoglobin A1C, s-crea = s-creatinine, s-urea = serum-urea.
P < .05.
P < .01.
Next, we performed correlation analyses between lipid parameters with different surrogate kidney-function markers (Table 5) and found a negative correlation between HDL-C and s-crea (P = .001), and a negative correlation between HDL-C and s-urea (P = .042). There was no significant correlation between HDL-C and proteinuria. LDL-C revealed no significant correlation with urinary or renal parameters. Lp(a) correlated positively with s-crea (P = .015) and the opposite with eGFR (P = .031) (Fig. 3). Moreover, Lp(a) correlated positively with the tp/crea ratio (P = .022) (Fig. 3). TC correlated negatively with s-crea (P = .044). Our TAG analysis revealed no significant correlations with renal/urinary markers. BMI correlated negatively with HDL-C (P = .001) and positively with TAG (P = .002). Furthermore, BMI correlated positively with s-crea (P = .003) and s-urea (P = .001).
Table 5.
Correlation coefficients between renal parameters and BMI with lipid parameters.
| Urinary albumin/creatinine | Urinary total protein/g creatinine | s-Creatinine | eGFR | s-Urea | BMI | ||
|---|---|---|---|---|---|---|---|
| HDL | r | −.062 | −.010 | −.226** | .037 | −.145* | −.371** |
| LDL | r | .063 | .049 | −.079 | −.003 | .064 | .041 |
| Lp(a) | r | .068 | .172* | .192* | −.143* | .060 | −.071 |
| Total cholesterol | r | .099 | .085 | −.141* | .046 | .016 | −.001 |
| Triglycerides | r | .111 | .021 | .133 | −.053 | .015 | .232** |
Values indicate correlations; correlations were calculated by Spearman; r = correlation coefficient.
BMI = body mass index, eGFR = electronic glomerular filtration rate, HDL = high density lipoprotein, LDL = low density lipoprotein, Lp(a) = lipoprotein (a), s-crea = s-creatinine, s-urea = serum-urea.
P < .05.
P < .01.
Figure 3.
Lp(a) correlates with increased proteinuria and loss of eGFR. Correlation graphs between Lp(a) with total proteinuria and eGFR in all participants. eGFR = electronic glomerular filtration rate, Lp(a) = lipoprotein (a).
Based on the previous results, we performed a multiple linear regression model with s-crea, eGFR, and urinary total protein/g creatinine ratio as respective dependent variables and age, HbA1c, BMI, systolic blood pressure, and lipid parameters as independent variables. All models themselves were statistically significant (P < .001). We found that Lp(a) values significantly predicted eGFR (standardized β = −0.183; P = .007) and urinary total protein/g creatinine ratio (standardized β = 0.213; P = .003), while no significant prediction on s-crea was observed (standardized β = 0.1; P = .084). Age was the only variable that significantly predicted all 3 independent variables and had the highest standardized regression coefficients for the prediction of eGFR and s-crea (eGFR: standardized β = −0.568, P < .001; s-crea: standardized β = 0.296, P = <.001; urinary total protein/g creatinine ratio: standardized β = 0.176, P = .011). All other lipid parameters did not significantly predict one of the independent variables.
4. Discussion
In the present study we aimed to investigate the association between dyslipidemia and its therapy with renal dysfunction. We therefore compared patients diagnosed with prevalent dyslipidemia based on their current treatment modality (lipid apheresis, drug treatment, and untreated). Because of its huge impact on the lipid profile, patients with diabetes were extracted from the aforementioned groups and examined separately.[13]
Patients in the LA and MG had lower levels of LDL-C, HDL-C, and TC, while Lp(a) levels were higher than in the untreated CG. With respect to apparently unaffected Lp(a) values under lipid apheresis, an earlier potential impact must be considered, since we did not know the initial Lp(a) levels before starting lipid apheresis. Significant reduced Lp(a) has been described in studies involving longitudinal data acquisition.[12]
We noted no significant correlation of LDL-C with renal or urinary parameters. However, in literature there is evidence of increased LDL-C and a loss of renal function correlating positively with albuminuria.[14] It should be taken into account that the initial LDL-C values before treatment initiation of the DG and MG are not known and a possible significant correlation with renal function cannot be excluded at this time-point. We found that HDL-C correlated negatively with the renal parameters s-crea and s-urea, but observed no significant correlation between HDL-C and proteinuria. There are conflicting findings in the literature regarding the above-mentioned association between HDL-C and elevated renal retention parameters. Previous results from a study by Schaeffner et al[1] are consistent with ours. However, Muntner et al reported no association between HDL-C and a loss of renal function[3] (a finding in line with the general protective cardiovascular effect attributed to HDL-C).[15] The important change in HDL-C’s protective characteristics in chronic renal failure, as reported by Zewinger et al, should be taken into account.[8] TC has shown a negative correlation with s-crea. This has also been reported in other studies.[16] TG revealed no significant correlation with renal parameters in our study.
Lipoprotein (a) correlated positively with higher urinary tp/g creatinine values in our cohort. It also revealed a positive correlation with s-crea, accompanied by a negative correlation with eGFR. Furthermore, multiple linear regression analysis revealed that Lp(a) values had the strongest predictive value on urinary total protein/g creatinine ratio and also significantly predicted eGFR. None of the other investigated lipid parameters showed significant prediction for kidney function related variables. Lp(a) thus seems to have a strong association with renal impairment. However, it should be noted that we recorded a significant overlap of Lp(a) levels between the study groups, which could result in reduced discriminative ability for patient populations. As discussed above, longitudinal assessment of Lp(a) values in future studies will be crucial to decipher its role in the progression of renal dysfunction.
Tada et al carried out similar investigations and reported an association between high Lp(a) levels and the incidence of chronic renal impairment independently of risk factors such as age, gender, BMI, blood pressure, diabetes, and LDL-C levels. However, a correlation with urinary protein has not been assessed.[17,18] The “lipid nephrotoxicity” effect comprises the progressive loss of renal function caused by dyslipidemia. It describes the accumulation of lipid metabolites and their final lipoproteins in tissues and organs resulting in organ malfunction and therefore also renal dysfunction.[4] Furthermore, proinflammatory processes, as well as the micro- and macrovascular damage caused by Lp(a) could have an impact on the loss of renal function.[19,20] Lp(a) levels are genetically determined and are not significantly influenced by lifestyle. Conventional lipid lowering drugs like statins are effective in reducing LDL-C, but have limited effects on Lp(a). More effective drugs like the antisense oligonucleotide (pelacarsen) and a small interfering ribonucleic acid (olpasiran) are currently examined by ongoing clinical trials. But at the moment, lipid apheresis is considered to be the only approach, that can effectively lower Lp(a) levels.[21]
Besides diabetes mellitus, other comorbidities like liver disease, thyroid disorders, and uncontrolled hypertension are known to affect the lipid profile. We measured TSH as a surrogate for thyroid function and assessed the intake of L-thyroxin. Overall, TSH levels were within the normal range in the vast majority of patients and balanced between the groups. Hypertension was diagnosed in most participants in the LA, MG, and DG and also in around 50% in the CG. Systolic blood pressure was slightly higher in CG, which could be explained by the high numbers of treated individuals in the other groups. With respect to liver disease, 2 cases of chronic inactive hepatitis B infection were known but no other entities of liver disease were diagnosed. Biochemical examination revealed that parameters of liver damage were within the normal range of almost all participants, only gGT values in the DG group were slightly increased. An association between increased gGT levels and diabetes mellitus has been described before.[22]
Altogether, we cannot formally exclude an impact of the described comorbidities on the lipid profile but it appears to be of less importance in our cohort.
An important finding of our study is the fact that proteinuria was not higher in patients undergoing lipid apheresis or drug therapy compared to the control group. It is well known that proteinuria is a strong risk factor for the worsening of kidney dysfunction.[23] Very few studies to date have examined a group of patients with treatment-required severe dyslipidemia in comparison to a group with treatment-naïve moderate dyslipidemia. In particular, the role played by proteinuria in association with kidney function has not yet been investigated in detail in this context. Especially elevated Lp(a) levels were particularly associated with decreased eGFR and increased proteinuria in our study and the effective treatment of dyslipidemia seems to be of importance to delay the loss of kidney function. Given the increasing burden of CKD and the accompanied financial pressure on healthcare systems, strategies to prevent or slow disease progression are of utmost general importance.[24] Moreover, it has been reported that patients undergoing percutaneous coronary intervention are at increased Lp(a)-associated risk in case of simultaneous decreased eGFR indicating a vicious interplay.[25] This highlights the importance to efficiently screen for Lp(a) elevation as it is still vastly underrecognized.[26] We therefore recommend that Lp(a) levels should be at least measured once in adults with high cardiovascular risk or a strong family history of premature atherothrombotic disease as also recommended by current European guidelines.[21] Today, lipid apheresis appears to be the only effective treatment option, but specific drugs are currently tested in clinical trials and could increase the applicability and reduce costs to lower Lp(a).
Furthermore, machine learning techniques will be useful in near future to get a more comprehensive knowledge about the role of Lp(a) and other lipid parameters in renal and cardiovascular disease severity and progression, like this is already available, for example, coronary artery disease.[27] These technologies also give the opportunity to overcome small sample size or unicentric study design, as this study contains, by using Shapley additive explanations. This could be applicable for interpreting complex relationships between lipid metabolism and renal outcomes and could help to combine several smaller studies to a large data set.[28]
4.1. Limitations
The study design of an unicentric cross-sectional study involves certain limitations. First of all, we collected all data at a single clinical center, meaning the generalizability is limited as the results can be influenced by specific local practices or patient demographics. The unicentric design also results in a limited number of participants. In addition, we recorded data only at study inclusion without follow-up assessments. Therefore, we could only report associations between lipid and renal parameters without the ability to establish causality between lipid disorders and kidney dysfunction. The individual treatment duration and therapy goals of the participants were not recorded. Furthermore, our study incorporated different age, gender, and BMI distributions between groups. To more clearly demonstrate the influence of these confounding variables on lipid metabolism and kidney function, matching groups based on these variables would be necessary. However, due to the limited number of participants after subdividing into the different study groups, effective matching was not possible. Moreover, the limited preexisting patient cohort in an ambulant single center hindered us to calculate an a priori power analysis. Larger multicentric studies with longitudinal data acquisition and balanced confounding effects like age, gender, and BMI across the study groups are required to draw more robust conclusions and establish causality.
In addition, no risk stratification was done concerning the European Society of Cardiology risk groups based on SCORE, meaning that individual LDL-C target values could not be assessed. We therefore focused on absolute LDL-C values in this study; risk stratification in European Society of Cardiology risk groups in conjunction with specifying target LDL-C value would be necessary to determine the relative risk.
Acknowledgments
The authors acknowledge statistical consultation from Dr Anna-Lena Friedel (University of Duisburg-Essen). The authors thank Ulrike Cuerten for language editing. We acknowledge support by the Open Access Publication Fund of the University of Duisburg-Essen.
Author contributions
Conceptualization: Walter Reinhardt.
Formal analysis: Nicole Lewandowski, Nils Mülling.
Investigation: Nicole Lewandowski.
Methodology: Kristina Boss.
Resources: Ralf Spitthöver, Andreas Kribben.
Supervision: Walter Reinhardt.
Validation: Kristina Boss.
Visualization: Nils Mülling.
Writing – original draft: Nicole Lewandowski, Nils Mülling.
Writing – review & editing: Walter Reinhardt, Ralf Spitthöver, Andreas Kribben, Kristina Boss.
Supplementary Material
Abbreviations:
- BMI
- body mass index
- CG
- control group
- DG
- diabetes group
- eGFR
- electronic glomerular filtration rate
- gGT
- gamma glutamyl transferase
- HDL-C
- high density lipoprotein-concentration
- IDDM
- insulin dependent diabetes mellitus
- LA
- lipid-apheresis treated group
- LDL-C
- low density lipoprotein-concentration
- Lp(a)
- lipoprotein (a)
- MG
- drug-treated group
- NIDDM
- non-insulin dependent diabetes mellitus
- s-crea
- serum-creatinine
- s-urea
- serum-urea
- TAGs
- triglycerides
- TC
- total cholesterol
- urinary tp
- urinary total protein
The authors have no funding to disclose.
All participants provided written informed consent before study inclusion. The study was reviewed and approved by the ethics committee of the University Duisburg-Essen (application 16-6883-BO).
AK received lecture fees, consulting fees, and travel fees from Alexion, Astellas, Astra Zeneca, Bayer Vital, GlaxoSmithKline, Janssen Cilag, Kyowa Kirin, Neovii, Novartis, Rhythm Pharmaceuticals, and Vifor Pharma. AK received research grants from BMBF, Cendres + Metaux SA, Charité, Exthera, Mirobio, MorphoSys, Numares, Quintiles, and Reata Pharmaceuticals. Remaining authors have no conflicts of interest to disclose.
All data generated or analyzed during this study are included in this published article [and its supplementary information files].
Supplemental Digital Content is available for this article.
How to cite this article: Lewandowski N, Reinhardt W, Spitthöver R, Kribben A, Boss K, Mülling N. Impact of severe dyslipidemia on renal function: A cross-sectional study. Medicine 2025;104:35(e44131).
Contributor Information
Nicole Lewandowski, Email: n.lewandowski5392@gmail.com.
Walter Reinhardt, Email: walter.reinhardt@uk-essen.de.
Ralf Spitthöver, Email: spitthoever@gmx.de.
Andreas Kribben, Email: andreas.kribben@uk-essen.de.
Kristina Boss, Email: kristina.boss@uk-essen.de.
References
- [1].Schaeffner ES, Kurth T, Curhan GC, et al. Cholesterol and the risk of renal dysfunction in apparently healthy men. J Am Soc Nephrol. 2003;14:2084–91. [DOI] [PubMed] [Google Scholar]
- [2].Fox CS, Larson MG, Leip EP, Culleton B, Wilson PW, Levy D. Predictors of new-onset kidney disease in a community-based population. JAMA. 2004;291:844–50. [DOI] [PubMed] [Google Scholar]
- [3].Muntner P, Coresh J, Smith JC, Eckfeldt J, Klag MJ. Plasma lipids and risk of developing renal dysfunction: the atherosclerosis risk in communities study. Kidney Int. 2000;58:293–301. [DOI] [PubMed] [Google Scholar]
- [4].Agrawal S, Zaritsky JJ, Fornoni A, Smoyer WE. Dyslipidaemia in nephrotic syndrome: mechanisms and treatment. Nat Rev Nephrol. 2018;14:57–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Mach F, Baigent C, Catapano AL, et al. 2019 ESC/EAS guidelines for the management of dyslipidaemias: lipid modification to reduce cardiovascular risk. Eur Heart J. 2020;41:111–88. [DOI] [PubMed] [Google Scholar]
- [6].Baigent C, Blackwell L, Emberson J, et al. Efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis of data from 170,000 participants in 26 randomised trials. Lancet. 2010;376:1670–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Kjeldsen EW, Thomassen JQ, Frikke-Schmidt R. HDL cholesterol concentrations and risk of atherosclerotic cardiovascular disease – insights from randomized clinical trials and human genetics. Biochim Biophys Acta Mol Cell Biol Lipids. 2022;1867:159063. [DOI] [PubMed] [Google Scholar]
- [8].Zewinger S, Speer T, Kleber ME, et al. HDL cholesterol is not associated with lower mortality in patients with kidney dysfunction. J Am Soc Nephrol. 2014;25:1073–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Kamstrup PR, Tybjaerg-Hansen A, Steffensen R, Nordestgaard BG. Genetically elevated lipoprotein(a) and increased risk of myocardial infarction. JAMA. 2009;301:2331–9. [DOI] [PubMed] [Google Scholar]
- [10].Klingel R, Heibges A, Fassbender C. Lipoprotein apheresis for Lp(a)-hyperlipoproteinemia with progressive cardiovascular disease – additional particular aspects of the Pro(a)LiFe multicenter trial. Atheroscler Suppl. 2015;18:35–40. [DOI] [PubMed] [Google Scholar]
- [11].Leebmann J, Roeseler E, Julius U, et al. Lipoprotein apheresis in patients with maximally tolerated lipid-lowering therapy, lipoprotein(a)-hyperlipoproteinemia, and progressive cardiovascular disease: prospective observational multicenter study. Circulation. 2013;128:2567–76. [DOI] [PubMed] [Google Scholar]
- [12].Heigl F, Hettich R, Lotz N, et al. Efficacy, safety, and tolerability of long-term lipoprotein apheresis in patients with LDL- or Lp(a) hyperlipoproteinemia: findings gathered from more than 36,000 treatments at one center in Germany. Atheroscler Suppl. 2015;18:154–62. [DOI] [PubMed] [Google Scholar]
- [13].Daniel MJ. Lipid management in patients with type 2 diabetes. Am Health Drug Benefits. 2011;4:312–22. [PMC free article] [PubMed] [Google Scholar]
- [14].de Boer IH, Astor BC, Kramer H, et al. Mild elevations of urine albumin excretion are associated with atherogenic lipoprotein abnormalities in the Multi-Ethnic Study of Atherosclerosis (MESA). Atherosclerosis. 2008;197:407–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Murphy AJ, Woollard KJ, Hoang A, et al. High-density lipoprotein reduces the human monocyte inflammatory response. Arterioscler Thromb Vasc Biol. 2008;28:2071–7. [DOI] [PubMed] [Google Scholar]
- [16].Bagheri B, Radmard N, Faghani-Makrani A, Rasouli M. Serum creatinine and occurrence and severity of coronary artery disease. Med Arch. 2019;73:154–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Tada H, Yamagami K, Nishikawa T, et al. Lipoprotein (a) and the risk of chronic kidney disease in hospitalized Japanese patients. Intern Med. 2020;59:1705–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Xuan L, Wang T, Dai H, et al. Serum lipoprotein (a) associates with a higher risk of reduced renal function: a prospective investigation. J Lipid Res. 2020;61:1320–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Ma KL, Gong TK, Hu ZB, et al. Lipoprotein(a) accelerated the progression of atherosclerosis in patients with end-stage renal disease. BMC Nephrol. 2018;19:192. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].van der Valk FM, Bekkering S, Kroon J, et al. Oxidized phospholipids on lipoprotein(a) elicit arterial wall inflammation and an inflammatory monocyte response in humans. Circulation. 2016;134:611–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Pirillo A, Casula M, Catapano AL. European guidelines for the treatment of dyslipidaemias: new concepts and future challenges. Pharmacol Res. 2023;196:106936. [DOI] [PubMed] [Google Scholar]
- [22].Sabanayagam C, Shankar A, Li J, Pollard C, Ducatman A. Serum gamma-glutamyl transferase level and diabetes mellitus among US adults. Eur J Epidemiol. 2009;24:369–73. [DOI] [PubMed] [Google Scholar]
- [23].Webster AC, Nagler EV, Morton RL, Masson P. Chronic kidney disease. Lancet. 2017;389:1238–52. [DOI] [PubMed] [Google Scholar]
- [24].Chadban S, Arici M, Power A, et al. Projecting the economic burden of chronic kidney disease at the patient level (inside CKD): a microsimulation modelling study. EClinicalMedicine. 2024;72:102615. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Zeng G, Zhu P, Yuan D, et al. Renal function alters the association of lipoprotein(a) with cardiovascular outcomes in patients undergoing percutaneous coronary intervention: a prospective cohort study. Clin Kidney J. 2024;17:sfae032. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Rendler J, Murphy M, Yeang C. Lipoprotein(a) is a prevalent yet vastly underrecognized risk factor for cardiovascular disease. Health Care Curr Rev. 2024;12:397. [PMC free article] [PubMed] [Google Scholar]
- [27].Huang AA, Huang SY. Use of machine learning to identify risk factors for coronary artery disease. PLoS One. 2023;18:e0284103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Huang AA, Huang SY. Increasing transparency in machine learning through bootstrap simulation and shapely additive explanations. PLoS One. 2023;18:e0281922. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.