Serum ferritin associated with atherogenic lipid profiles in a high-altitude living general population

Introduction

Globally, cardiovascular disease (CVD) ranks as the top cause of death (Arnett et al., 2019; Visseren et al., 2021). Low-density lipoprotein cholesterol (LDL-C) is a recognized independent factor contributing to CVD risk and a primary target for CVD prevention (Mach et al., 2020). A subset of LDL, known as small dense LDL (sdLDL), holds a particularly strong association with the risk of atherosclerotic CVD (ASCVD) (Krauss, 2022; Yanai et al., 2022). Notably, lipoprotein a (Lp(a)) has been proven as a causal risk factor for ASCVD, independent of LDL-C (Reyes-Soffer et al., 2022). Additionally, it has been found that triglycerides (TG) and triglyceride-rich lipoproteins (TGRL) play an important in the residual risk realated to ASCVD (Dhindsa et al., 2020).

Inflammation has various maladaptive functions that lead to the progression and instability of ASCVD, and high-sensitivity C-reactive protein (hs-CRP) serve as a main clinical indicator of inflammation (Lawler et al., 2021). Furthermore, serum amyloid A (SAA) and serum ferritin (SF) are also important in acute and chronic inflammation, and are used as additional markers of inflammatory conditions (Kell & Pretorius, 2014; Zhang et al., 2019).

SF, apart from being an inflammation marker, is also vital in the iron delivery system and is gauged as an indicator of iron status (Kell & Pretorius, 2014; Lopez et al., 2016). People who live at high altitudes for a long time are constantly in a hypoxic environment. Erythrocytosis is an important adaptation mechanism that requires substantially higher amounts of iron (Gassmann & Muckenthaler, 2015). A study on young Swiss male whose residential altitudes range from 200 to 2,000 m revealed that for every 300 m increase in residential altitude, the levels of hemoglobin (Hb) and SF increased significantly (Staub et al., 2020). However, a study conducted in Han Chinese males showed that SF levels of participants living at high altitudes did not increase compared with those of sea level residents (Liu et al., 2018). The iron stores of high-altitude residents were distinct among different populations. In healthy people, iron stores were in the physiological range in spite of the increased need for erythropoiesis. However, for vulnerable groups living at a high altitude with a greater demand for iron, their iron stores were not likely to be quickly replenished as in normal situations (Böning et al., 2004; Burke et al., 2017; Cook et al., 2005; Tufts et al., 1985).

Preliminary studies conducted in different populations, including Spanish (González et al., 2006), American (Li et al., 2023), Qatar (Al Akl et al., 2021), and Chinese (Li et al., 2017; Zhou, Liu & Yuan, 2021) adults as well as Korean (Kim et al., 2016) and Chinese (Zhu et al., 2019) adolescents, have demonstrated that SF was significantly associated with dyslipidemia. Nevertheless, the association between SF and atherogenic lipid profiles in high-altitude living populations was unclear. Given the remarkable effect of high altitude on iron stores and SF levels, studies investigating the correlations between SF and dyslipidemia specifically in high-altitude residents are necessary. To fill this knowledge gap, we investigate for the first time the association between SF and atherogenic lipid profiles in a general population of Tajik adults living on the Pamirs Plateau of China.

Materials & Methods

Results

Correlation between SF and atherogenic lipid profiles

SF was significantly correlated with most of the lipids, including TC (Beta = 0.225, P < 0.001), LDL-C (Beta = 0.197, P < 0.001), HDL-C (Beta = −0.218, P < 0.001), TG (Beta = 0.332, P < 0.001), and sdLDL-C (Beta = 0.316, P < 0.001). An exception was observed with Lp(a), which showed no significant correlation (Beta = 0.018, P = 0.475) (Fig. 1).

Association of SF concentration with different types of dyslipidemia

In the male population, higher SF was positively correlated with high TC (OR: 1.413, 95% CI [1.010–1.978]), high TG (OR: 1.602, 95% CI [1.299–1.976]), and high sdLDL-C (OR: 1.631, 95% CI [1.370–1.942]). After adjusting for age (model 1), SF maintained a significant correlation with high TG (OR: 1.610, 95% CI [1.303–1.990]) and high sdLDL-C (OR: 1.610, 95% CI [1.353–1.917]). In model 2 (adjusting for WC, BMI, and BP based on model 1) and model 3 (adjusting for smoking, drinking, and exercise based on model 2), SF concentration remained significantly correlated with high sdLDL-C (OR: 1.470, 95% CI [1.224–1.766]) and high TG (OR: 1.382, 95% CI [1.100–1.737]) (Fig. 2).

In the female population, higher SF was positively correlated with high TC (OR: 1.461, 95% CI [1.061–2.014]), high LDL-C (OR: 2.104, 95% CI [1.481–2.990]), low HDL-C (OR: 1.447, 95% CI [1.195–1.752]), high TG (OR: 2.106, 95% CI [1.454–3.050]), and high sdLDL-C (OR: 2.000, 95% CI [1.589–2.516]). In model 1, higher SF was significantly correlated with high TG (OR: 1.970, 95% CI [1.260–3.079]), low HDL-C (OR: 1.416, 95% CI [1.120–1.790]), and high sdLDL-C (OR: 1.424, 95% CI [1.087–1.865]). In model 2 and model 3, higher SF concentration remained significantly correlated with low HDL-C (OR: 1.348, 95% CI [1.061–1.711]) and high TG (OR: 1.677, 95% CI [1.070–2.628]) (Fig. 3).

Table 2:

Sex specific correlations between serum ferritin concentrations and lipid profiles stratified by hs-CRP and SAA levels.

	TC (mmol/L)		LDL-C (mmol/L)		HDL-C (mmol/L)		TG (mmol/L)		Lp(a) (mg/L)		sdLDL-C (mg/L)
	Beta	P	Beta	P	Beta	P	Beta	P	Beta	P	Beta	P
Men
hs-CRP (mg/L)
Q1	0.245	<0.001	0.066	0.333	−0.055	0.427	0.334	<0.001	0.007	0.922	0.154	0.024
Q2	0.290	<0.001	0.231	0.001	0.004	0.955	0.168	0.013	−0.057	0.405	0.309	<0.001
Q3	0.139	0.044	0.052	0.454	0.060	0.383	0.185	0.007	−0.061	0.381	0.177	0.010
Q4	0.186	0.005	0.155	0.021	−0.069	0.306	0.222	0.001	0.134	0.050	0.231	0.001
SAA (mg/L)
Q1	0.295	<0.001	0.250	<0.001	−0.003	0.965	0.212	0.003	0.025	0.725	0.289	<0.001
Q2	0.133	0.045	0.089	0.180	−0.087	0.192	0.202	0.002	0.004	0.954	0.201	0.002
Q3	0.194	0.005	−0.011	0.875	−0.110	0.110	0.375	<0.001	−0.110	0.114	0.151	0.028
Q4	0.249	<0.001	0.176	0.008	0.078	0.243	0.164	0.013	0.062	0.361	0.241	<0.001
Women
hs-CRP (mg/L)
Q1	0.324	<0.001	0.349	<0.001	−0.254	<0.001	0.361	<0.001	−0.072	0.307	0.369	<0.001
Q2	0.307	<0.001	0.327	<0.001	−0.193	0.005	0.275	<0.001	0.091	0.193	0.297	<0.001
Q3	0.373	<0.001	0.409	<0.001	−0.216	0.002	0.275	<0.001	0.024	0.729	0.414	<0.001
Q4	0.279	<0.001	0.313	<0.001	−0.219	0.001	0.337	<0.001	−0.043	0.546	0.341	<0.001
SAA (mg/L)
Q1	0.359	<0.001	0.355	<0.001	−0.247	0.001	0.381	<0.001	−0.072	0.362	0.453	<0.001
Q2	0.375	<0.001	0.390	<0.001	−0.288	<0.001	0.376	<0.001	0.115	0.087	0.410	<0.001
Q3	0.340	<0.001	0.369	<0.001	−0.201	0.002	0.299	<0.001	−0.006	0.930	0.329	<0.001
Q4	0.188	0.006	0.263	<0.001	−0.275	<0.001	0.248	<0.001	−0.035	0.618	0.271	<0.001

DOI: 10.7717/peerj.19104/table-2

Notes:

hs-CRP

high-sensitivity C-reactive protein

SAA

serum amyloid A

TC

total cholesterol

LDL-C

low-density lipoprotein cholesterol

HDL-C

high-density lipoprotein cholesterol

TG

triglycerides

Lp(a)

lipoprotein a

sdLDL-C

small dense LDL-C

Q1

first quartile

Q2

second quartile

Q3

third quartile

Q4

fourth quartile

Beta

standardized regression coefficients

P

p values

Subgroup analysis for young (18–35 years) and middle-aged (≥45 years) participants

In the young participants group, SF was significantly correlated with TC (Beta = 0.192, P < 0.001), LDL-C (Beta = 0.168, P < 0.001), HDL-C (Beta = −0.313, P < 0.001), TG (Beta = 0.374, P < 0.001), and sdLDL-C (Beta = 0.366, P < 0.001). There was no correlation between SF and Lp(a) (Beta = −0.008, P = 0.845) (Fig. S1). In the middle-aged participants group, SF was significantly correlated with TC (Beta = 0.162, P < 0.001), LDL-C (Beta = 0.110, P = 0.005), HDL-C (Beta = −0.162, P < 0.001), TG (Beta = 0.299, P < 0.001), and sdLDL-C (Beta = 0.214, P < 0.001). There was no correlation between SF and Lp(a) (Beta = −0.001, P = 0.977) (Fig. S2).

In the young population, higher SF quartile was positively correlated with high TG (OR: 1.702, 95% CI [1.203–2.407]) and high sdLDL-C (OR: 1.434, 95% CI [1.061–1.937]). There was no correlation between SF quartile and high TC (OR: 1.167, 95% CI [0.677–2.011]), high LDL-C (OR: 0.693, 95% CI [0.295–1.624]), low HDL-C (OR: 1.123 95% CI [0.930–1.356]), and high Lp(a) (OR: 0.927, 95% CI [0.724–1.167]) after adjusting for gender, WC, BMI, SBP, smoking, drinking, and exercise (Fig. 4A). In the middle-aged participants group, higher SF quartile was positively correlated with high LDL-C (OR: 1.459, 95% CI [1.042–2.044]), high TG (OR: 1.325, 95% CI [1.003–1.749]), and high sdLDL-C (OR: 1.502, 95% CI [1.218–1.852]). There was no correlation between SF quartile and high TC (OR: 1.307, 95% CI [0.926–1.844]), low HDL-C (OR: 1.016 95% CI [0.853–1.209]), or high Lp(a) (OR: 1.007, 95% CI [0.891–1.302]) (Fig. 4B).

Discussion

In this study, it was found that SF concentration was significantly higher in men compared to women. SF exhibited positively correlations with TC, LDL-C, TG, and sdLDL-C. Conversely, it displayed a negatively correlation with HDL-C. There was no significant correlation with Lp(a). Under different inflammatory conditions, SF was positively correlated with LDL-C and negatively correlated with HDL-C in women. SF was positively correlated with TC, TG, and sdLDL-C in both men and women. Higher SF levels were associated with high TC, high TG, and high sdLDL-C in men, and almost all kinds of dyslipidemia except for high Lp(a) in women. Even after adjusting for confounders, higher SF levels remained significantly associated with high TG in both men and women. The correlations between SF and TC, LDL-C, HDL-C, TG, and sdLDL-C were still significant in both young and middle-aged subgroups.

People who have long-term residence at high altitudes are in a hypoxic environment. Hypoxemia is sensed by prolyl hydroxylases (PHDs), which stabilize hypoxia-inducible factor (HIF). HIF can promote the expression of erythropoietin (Epo), increasing red blood cell production (Gassmann & Muckenthaler, 2015). Epo, hepcidin, and erythroferrone are involved in coordinating the liver, bone marrow, and small intestine, as well as altering iron acquisition. Therefore, altitude is closely related to iron metabolism. In this study, the mean Hb levels of high-altitude Tajik were up to 168.17 g/L in males and 149.28 g/l in females, indicating a high-altitude adaptation. Several studies conducted in different populations living at high altitudes showed that iron stores were distinct in healthy individuals and vulnerable populations with increased iron demand (Böning et al., 2004; Burke et al., 2017; Cook et al., 2005; Tufts et al., 1985). We reported that the median SF of Tajik males and females were 200.20 ng/mL and 42.30 ng/mL, respectively. Compared to the study conducted on Bolivian adults living at 3,500 m of altitude or above, the SF levels of male Tajik were much higher, but the SF levels of female Tajiks were similar (Beall et al., 1990). Regrettably, we have no data on SF in low-altitude Tajik populations since most Tajiks live on the high-altitude Pamirs Plateau of China, which prevented us from investigating the effects of high altitude on iron store.

Previous research indicated a consistent pattern regarding the relationship between SF and lipid profiles, showing positive correlations with TC, LDL-C, and TG, while revealing negative correlation with HDL-C levels in populations from the United States (Li et al., 2023) and Middle East (Al Akl et al., 2021). In this study conducted among high-altitude Tajik individuals, these trends were similarly observed. Notably, serum Lp(a) levels are mainly determined by genetic factors (Kronenberg et al., 2022; Reyes-Soffer et al., 2022), and the prevalence of high Lp(a) remained unassociated with SF in both men and women in this study. The exact way in which SF concentrations are related to dyslipidemia has not been fully grasped yet. Numerous studies have proven that SF was associated with other CVD risk factors, such as diabetes (Sun et al., 2013), hypertension (Piperno et al., 2002), elevated fasting insulin and blood glucose (Tuomainen et al., 1997), central adiposity (Gillum, 2001), and metabolic syndrome (Jehn, Clark & Guallar, 2004; Suárez-Ortegón et al., 2018), suggesting that SF may play a significant role in metabolism. SF could lose its iron and the unliganded iron has the capacity to catalyze hydroxyl radical formation, leading to oxidative tissue damage (Kell & Pretorius, 2014; Wolff, 1993). Studies on rats showed that dietary iron restriction can lead to decreased TG concentrations and reduced levels of lipid peroxidation levels (Vargas-Vargas et al., 2022). However, iron overload may induce elevated TG levels through several mechanisms, including enhanced hepatic oxidative stress, decreased fatty acid beta-oxidation, and the promoted hepatic lipid secretion by increasing the expression of apoB-100 and microsomal triglyceride transfer protein (Silva et al., 2015).

SF, apart from being a sign of iron storage, serves as a crucial marker for inflammatory diseases too, since it is mainly a product leaked from damaged cells (Kell & Pretorius, 2014). Therefore, is the relationship between SF and lipid profiles in high-altitude residents associated with its role as an inflammation marker? A previous study conducted in the general Chinese population found that the association between SF and lipid parameters persisted after adjusting for CRP levels, suggesting that low-grade inflammation may not be the sole explanation for the link between SF levels and dyslipidemia (Li et al., 2017). In this study, individuals with chronic inflammatory diseases were excluded. Notably, SF showed no significant correlation with two other inflammation markers, hsCRP and SAA. Moreover, SF remained significantly associated with TC, TG, and sdLDL-C across subgroups stratified by varying levels of hs-CRP and SAA. These findings suggest that the relationship between SF and lipid levels cannot be attributed to its role as an inflammation marker.

There seemed to be a gender discrepancy in the association between SF and lipids levels or dyslipidemia. A study conducted on Korean adolescents found that SF was positively associated with TC, LDL-C, and TG levels, and negatively associated with HDL-C levels in women. However, in men, the association was observed only with HDL-C levels (Kim et al., 2016). This result differed from a study conducted in US adults, where only females showed a clearly increasing tendency for high-TC and high-LDL-C. In this study, the relationship between SF and lipid levels appeared to be stronger in women, which aligned with the findings from the US study. In women, five out of six lipid parameters were correlated with SF across varying levels of inflammation. In contrast, only three out of six lipid parameters were significantly associated with SF in men. Regarding dyslipidemia, nearly all types of dyslipidemia, except for high Lp(a), were associated with higher SF in women. However, after adjusting for confounders, SF was only associated with high TG in both men and women, indicating that the gender discrepancy in the association between SF and lipids might be attributed to differences in the distributions of confounders between men and women.

There are a number of limitations in this study which need to be taken into account. First, because of its cross-sectional nature, the study was unable to determine a causal connection between SF and lipid profiles. Prospective cohort studies would be necessary to determine a causal relationship. Second, although this study was conducted in a single ethnic group with similar eating habits, dietary information was not included in the study, making it challenging to adjust for dietary factors (especially iron intake) in the analysis of the relationship between SF and lipids. Third, the menstrual status of women was unknown, which might have introduced potential bias to the results.

Despite these limitations, a key strength of this study lies in its novel investigation of the association between SF and lipid profiles under varying inflammatory conditions in high-altitude living residents. Moreover, as far as we know, this is the first study to look into the relation between SF and Lp(a) as well as sdLDL-C. These findings help us better understand the interplay between SF and lipids, providing useful perspectives and create a basis for more researches in the future.

In conclusion, SF was positively correlated with TC, LDL-C, TG, and sdLDL-C; negatively correlated with HDL-C; and not significantly correlated with Lp(a). Even after adjusting for confounders, higher SF levels were associated with high TG in both genders. The relationship between SF and lipid levels cannot be attributed to its role as an inflammation marker.

Supplemental Information