In this large population-based study, the association between elevated tHcy levels and higher prevalence of albuminuria was observed independently of GFR, reflecting the fact that this association is not simply the consequence of reduced renal function. This result is consistent with the hypothesis that homocysteine may cause renal damage. We extend previous population-based findings [1, 10, 11] to the age group 35-40 years. Also, the association of hyperhomocysteinemia with albuminuria was independent of hypertension, type 2 diabetes and uric acid levels. Furthermore, hyperhomocysteinemia was associated with a 2-fold higher risk of albuminuria, which was of similar magnitude to the risk of albuminuria associated with hypertension or type 2 diabetes. This result strongly suggests that hyperhomocysteinemia is an independent marker of renal dysfunction and is in line with the role of homocysteine as potential atherogenic agent and cardiovascular risk factor .
We observed strong associations between homocysteine levels and MTHFR polymorphisms, as previously reported [29–31]. This suggests that MTHFR genotypes related to higher homocysteine levels are associated with albuminuria in the general population, in particular in non-diabetic subjects. Because of the random assortment of alleles at the time of gamete formation, association between MTHFR genotypes and tHcy levels, or albuminuria, should not be biased by reverse-causality or major confounding factors. These results therefore suggest that homocysteine might be causally involved in renal damage. MTHFR polymorphisms have been shown to be associated with increased mortality and morbidity in healthy and CKD populations [29–31]. Some studies in diabetic patients found the C677T polymorphism (rs1801133) to be associated with renal damage [32–36], but others failed to find such association [37, 38]. In this study, rs1321073 was significantly associated with the risk of albuminuria in the general population, independently of major confounding factors. However, this latter association was not significant in type 2 diabetic subjects, in whom no association with tHcy was observed. First of all, the number of diabetics was small. Second, since diabetes itself represents a major cause of albuminuria, the effect of homocysteine on renal damage might not be easily detectable in this subgroup, as reflected by the conflicting results reported in studies restricted to diabetic patients [32–38]. Results for the nonsynonymous rs1801133 (C677T) SNP, which is associated with reduced MTHFR activity , were quite similar. Although associated with tHcy levels, rs9651118 was not significantly associated with albuminuria, but the effect was nevertheless in the same direction. As each SNP only explained a tiny fraction of homocysteine variance, this study might have reduced power to detect this association.
In line with previous reports, we found a positive association between SUA and tHcy . However, SUA did not modify the relationship between tHcy levels and albuminuria. This suggests that there is no strong synergistic or antagonistic effect between homocysteine and uric acid for their effect on renal damage, but rather that both factors act independently on albuminuria. Conversely, we observed an effect modification of uric acid on the known negative association between tHcy levels and GFR, independently of major confounding factors. We are not aware that the effect modification of SUA on the association between tHcy and renal function has been previously reported. Both hyperuricemia and hyperhomocysteinemia are common findings in patients with CKD. In prospective studies among CKD patients, baseline SUA and baseline tHcy levels appeared to predict overall mortality [4, 18]. However the link between tHcy, SUA and decreased renal function is unclear. A possible explanation is that antioxidants, like uric acid, raise in reaction to increased oxidative stress to counteract the harmful effect of reactive oxygen species induced by elevated homocysteine . Reactive oxygen species reduce glomerular blood flow and filtration rate by release of vasoconstrictors and, possibly by inactivation of NO .
Hyperhomocysteinemia is frequent in patients with CKD [1, 4]. In this study, tHcy was higher whenever renal function was decreased. This might result, in part, from the fact that impaired renal function reduces the renal clearance of homocysteine. This could also result, in part, from impaired intrarenal homocysteine metabolism, impaired extrarenal homocysteine metabolism or both. Also the prevalence of albuminuria was gradually higher with increasing tHcy levels, independently of glomerular filtration rate, since a similar trend was observed after stratifying for normal and decreased GFR levels. Several prospective studies have shown that hyperhomocysteinemia is an independent determinant of the development of CKD and albuminuria in the general population [11, 40]. Moreover, in a prospective study, kidney transplant recipients with elevated tHcy levels at baseline had a greater risk of death and kidney allograft loss . These findings further support the concept that tHcy is not merely a marker, but rather a predictor of CVD. CVD currently represents the main cause of death in CKD patients, with mortality rates three to 30 times higher than expected [6, 41]. As this high mortality is only partly explained by CVD risk factors, such as hypertension, diabetes, smoking and dyslipidemia, factors such as hyperhomocysteinemia and elevated uric acid levels may play a role.
Our study has several limitations. Unfortunately, we have no data on vitamin status, including folic acid, vitamin B6, and vitamin B12, but we could include data on reported vitamin B intake. However, several previous studies observed no significant interaction of vitamin B status on the association between tHcy and renal function or tHcy and cardiovascular risk [1, 42]. We used estimations of GFR by the MDRD formula in this study. As MDRD has not been validated in populations with normal GFR, we also conducted sensitivity analyses using GFR estimated by the Cockcroft-Gault formula, which led to similar results. The calculated GFR and albumin-to-creatinine ratio (ACR) were based on a single measurement, as in previous population-based studies [1, 10]. tHcy was also based on one single measure and is therefore insufficient to predict life-time exposure. Despite the small proportion of homocysteine variance explained, MTHFR polymorphisms may better reflect the life-time exposure to tHcy. However, our study also has several strengths. It is to our knowledge the largest population-based study exploring associations between homocysteine and renal function, hence with a wide external validity. A longitudinal follow-up of all participants in the CoLaus study is currently ongoing and shall provide essential data for trends over time of major cardiovascular risk factors according to baseline homocysteine levels.