Diabetic peripheral neuropathy (DPN) is the most common form of diabetic neuropathy [1
]. It is an important cause of foot ulceration and a major contributor to falls and fractures [2
]. Long-duration diabetes, old age, hyperglycemia, hypertension, dyslipidemia, obesity, alcohol, smoking, and insulin resistance are known risk factors for DPN [3
]. From a pathophysiologic point of view, oxidative stress is a key contributor to DPN [5
]. However, alpha-lipoic acid, a currently available antioxidant treatment, showed a clinically relevant effect on symptomatic DPN when administered by intravenous infusion [6
]. Furthermore, there is no approved preventive or curative treatment for DPN other than risk factor management. Therefore, the identification of additional modifiable factors is crucial for developing a new strategy to treat DPN.
Diet plays a critical role in the development of type 2 diabetes and its complications [7
]. The Mediterranean [9
] and vegetarian [10
] diets reduce diabetes risk. In addition, a low-carbohydrate, high-unsaturated and low-saturated fat diet improved glycemic control and cardiovascular disease risk factors [11
]. Several studies showed that the blood concentration of micronutrients was decreased in individuals with DPN. For example, vitamin D deficiency is associated with DPN [12
] and with painful DPN even after adjusting for confounding factors [13
]. Among metformin users, vitamin B12 deficiency was commonly detected and might have a detrimental effect on DPN [14
]. The blood levels of other vitamins B such as vitamin B1 and B6 [15
] were also associated with a high frequency of DPN. Therefore, assessing these micronutrients might be useful for the stratification of DPN risk [16
]. However, from the perspective of clinical practice, the measurement of the blood levels of micronutrients might not be feasible in terms of cost and complexity of the procedure. In this regard, nutritional assessment might be a solution to evaluate an individual’s risk of DPN.
Iron, a trace element, participates in a variety of cellular processes, including oxygen delivery, mitochondrial electron transport, DNA synthesis, and gene regulation [17
]. However, excess iron can generate oxidative stress and cause tissue damage [18
]. Interestingly, previous studies have shown an association between dietary iron intake and diabetes risk [19
]. Several rodent models of DPN have shown that iron deficiency rather than iron overload was associated with the risk of DPN [21
]. However, no study has evaluated the association between dietary iron intake and DPN in humans.
Polyunsaturated fatty acids (PUFA), especially, omega-3 PUFA, are antioxidants [24
]. Intake of PUFA and replacement of saturated fatty acids (SFA) with PUFA reduced the risk of type 2 diabetes [25
]. In addition, a high dietary PUFA intake was associated with a lower risk of DPN [27
]. Despite these findings, few studies have evaluated the association between these nutrients and DPN. Therefore, in this study, we examined the association of iron intake and of the ratio between iron intake and PUFA intake (iron/PUFA) with DPN in individuals with type 2 diabetes.
In this cross-sectional study, we observed that iron intake and iron intake relative to PUFA levels were higher in participants with DPN than in participants without DPN. Furthermore, the presence of DPN or the severity of DPN assessed by the MNSI-PE was positively associated with iron intake and the iron/PUFA ratio.
Prospective cohort studies of the Japanese and Chinese populations reported that iron intake was associated with an increased risk of diabetes [20
]. In a case–control study of Europids with type 2 diabetes, a hemochromatosis-causing mutation C282Y was associated with a higher risk of diabetic retinopathy [33
]. In a rat model, iron caused renal tubular injury due to the formation of free hydroxyl radicals [34
]. A prospective intervention study revealed that a low-iron diet delayed the progression of diabetic nephropathy [35
]. In regard to neuropathy, an in vitro study demonstrated that iron overload aggravated the oxidative stress injury of neurons in the presence of high glucose concentrations [36
]. In our study, iron intake was associated with DPN, and this association was no longer significant after adjustment for HOMA-IR or HOMA-B. Therefore, insulin resistance and pancreatic beta cell dysfunction might be an important factor promoting the association between iron and DPN.
Oxidative stress causes pancreatic beta cell dysfunction, insulin resistance, and diabetic complications [37
]. Meanwhile, iron can generate oxidative stress by the formation of hydroxyl radicals through the Fenton reaction [39
]. Therefore, we need to measure reactive oxygen species (ROS) to understand the possible mechanism underlying the role of iron in DPN. However, the measurement of ROS is very tricky due to their reactivity and unstable properties [40
]. Therefore, it might be more reasonable to measure oxidation target products of ROS. Previous cross-sectional studies reported that serum NSE and nitrotyrosine levels, which are oxidation target products of ROS, were closely associated with DPN [41
]. In this background, we tested whether these indices were associated with DPN and useful for identifying DPN. However, we did not observe any differences in the levels of these biomarkers between individuals with DPN and those without DPN. This negative result might relate to the characteristics of the study subjects. We enrolled participants with relatively well-controlled type 2 diabetes, in contrast to the earlier study which enrolled subjects with both type 1 and type 2 diabetes [41
], and with a more severe degree of hyperglycemia; the mean HbA1c
was up to 9.4% [42
]. In fact, there are other oxidative markers of ROS-induced modifications of lipids, proteins, and DNA or RNA that we need to measure further [43
]. For example, serum malondialdehyde and urinary 8-hydroxy-2′-deoxy-guanosine might be good candidates, because their levels were shown to be higher in individuals with diabetic nephropathy [44
] and in individuals with diabetic microvascular complications [45
]. However, we did not have available samples to test these molecules, which is one of the limitations of our study.
Hepcidin is an established master regulator of iron metabolism and an index of the iron pool in the body [46
] that predicted the progression of diabetic nephropathy, one of the microvascular complications of type 2 diabetes [47
]. However, in the present study, we did not observe differences in the levels of serum transferrin, ferritin, iron, and hepcidin linked to the presence of DPN. Therefore, we cautiously infer that a dietary pattern including high-iron-containing food might be a more important risk factor for DPN than the actual amount of iron. In addition, there are two types of dietary iron: heme and non-heme iron. Heme iron is present in red meat, poultry, and seafood, while non-heme iron is present in both plant and animal foods. Heme iron contributes 10–15% of total iron intake, but because of its higher absorption, it can contribute over 40% of the total absorbed iron [48
]. As it is thought that the gut microbiota can influence the absorption capacity of iron [49
], it might be necessary to consider both the amount and the quality of iron intake, as well as gut environmental factors, to best assess iron absorption.
Our results are different from results of rodent models of DPN which showed that iron deficiency rather than iron overload was associated with the risk of DPN [21
]. It is hard to compare the results from human and rodent studies. In addition, cross-sectional studies and the intervention studies are different in terms of interpreting cause and effect. Even after consideration of the aforementioned points, we suggest that following might lead to different results in rodent models of DPN. First, in streptozotocin-diabetic rats, a single high dose of streptozotocin could induce nonspecific toxicity, which affects neurons directly [50
]. In that circumstance, iron deficiency might cause impairment of iron-containing repair enzymes. Second, in ob/ob
mice, iron deficiency can cause iron-deficiency anemia. The hemoglobin level in db/db
mice on a high-iron diet was 19.4 g/dL, while it was 10.7 g/dL in db/db
mice on a low-iron diet. This difference in hemoglobin might cause ischemia in the peripheral limbs. In contrast, considering that the individuals in our study did not have iron-deficiency anemia, iron deficiency might not have influenced our study results.
PUFA can be divided into two subclasses: omega-6 and omega-3. Omega-6 PUFA include linoleic acid and arachidonic acid [51
]. Omega-3 PUFA include alpha-linolenic acid, eicosapentaenoic acid, and docosahexaenoic acid [52
]. Omega-3 PUFA is an antioxidant able to produce a direct superoxide scavenging effect [24
] and an indirect reactive oxygen species reduction effect via upregulation of antioxidant molecules [53
]. Previous cross-sectional studies demonstrated that PUFA intake was associated with a lower odds ratio (OR) for the presence of diabetic retinopathy [54
] and that linolenic acid intake was associated with lower odds of peripheral neuropathy [27
]. A meta-analysis revealed that omega-3 fatty acid supplementation reduced the amount of proteinuria in individuals with type 2 diabetes [55
]. For the cardiovascular risk, a few randomized controlled trials [56
] and a meta-analysis [58
] did not show a benefit of omega-3 PUFA, but the Japan EPA Lipid Intervention Study (JELIS) [59
] and the Reduction of Cardiovascular Events with Icosapent Ethyl-Intervention Trial (REDUCE-IT) [60
] have shown a benefit of omega-3 PUFA [60
]. Until now, there is controversy about the role of omega-3 PUFA supplements for individuals with diabetes in the prevention of cardiovascular events [7
]. To our best knowledge, there is a lack of studies investigating the role of omega-3 PUFA supplements for DPN. Considering the high prevalence of DPN and the limited treatment options for DPN, it is valuable to investigate the association between omega-3 PUFA and the risk of DPN. In this study, we observed a lower trend of PUFA and omega-3 PUFA intake in individuals with DPN compared to those without DPN. Meanwhile, omega-6 PUFA is considered pro-inflammatory by some researchers because linoleic acid, the representative of omega-6 PUFA, is converted into arachidonic acid. In addition, a few studies suggested that omega-6 PUFA is related to chronic inflammatory diseases such as obesity, nonalcoholic fatty liver disease, cardiovascular disease [61
]. However, other studies showed that high consumption of omega-6 PUFA did not increase cardiovascular events [63
]. In this study, we observed a lower, but not significant, omega-6 PUFA intake in individuals with DPN. Considering the significant correlation between omega-6 PUFA intake and omega-3 PUFA intake in this study (r
= 0.713, p
< 0.001), it is not possible to interpret the results of omega-6 PUFA and omega-3 PUFA separately.
In light of the in vitro [36
] and animal model [65
] data, we postulated that increased body iron can damage neurons or Schwann cells via direct or indirect pathways. Considering the PUFA-related antioxidant effect observed in an iron-related, pro-oxidant environment, we calculated the iron/PUFA ratio and found that a higher iron/PUFA ratio was associated with a higher OR of DPN. This finding suggests that the ratio of iron to PUFA might be an important marker of DPN (Figure 1
) and can be used as an indicator to screen for or prevent DPN in individuals with type 2 diabetes. In addition, even though the ratio iron/omega-6 PUFA, rather than the ratio iron/omega-3 PUFA, showed a statistically significant association with DPN after adjusting for confounders, we need to be cautious in interpreting these data. A relatively small amount of omega-3 PUFA compared with omega-6 PUFA might bring about these non-significant results.
Our study has several limitations. First, because of its cross-sectional nature, we could not establish a causal relationship. Second, our study was based on a relatively small sample size, which may have affected the assessment of significant differences in known risk factors for DPN, such as age, BMI, and diabetes duration. Studies with a larger sample size or a prospective or intervention study would be of interest to confirm or reassess these findings. Third, neurophysiologic studies were not used to confirm the DPN diagnosis. Fourth, the Korean Food Composition Table does not contain data regarding haem iron content, so we could not analyze the intake of heme iron and non-heme iron separately. Lastly, among various oxidative stress markers, we measured only NSE and nitrotyrosine levels, which were comparable between groups.
Despite these limitations, this study has several strengths. First, we obtained dietary nutrient intake estimates with the use of three-day food records. In addition, we controlled for a number of dietary and nondietary covariates to reduce possible confounding effects. Above all, this is the first study to examine the association between dietary iron intake and DPN. In addition, we suggest the iron/PUFA ratio as a new index associated with DPN.