Abstract
Background:
Circulating lipids have been implicated as important modulators of immune response and altered lipid levels correlate with the severity of infection. However, long-term prognostic implications of lipid levels regarding future infection risk remain unclear. The current project aims to explore whether baseline lipid levels are associated with risk of future serious infection, measured by hospitalization for pneumonia.
Methods:
A retrospective analysis was performed in 13,478 participants selected from the Atherosclerosis Risk in Communities (ARIC) study, a large community-based longitudinal cohort in the United States with a median follow up time of >20 years. First incident of hospitalization for pneumonia was identified through hospital discharge records. Cox proportional hazard models were used to assess the association of baseline major lipid levels (total cholesterol, LDL-C, HDL-C, triglycerides) with time to first pneumonia hospitalization.
Results:
A total of 1969 (14.61%) participants had a pneumonia hospitalization during a median follow up time of 21.5 years. The hazard ratio (HR) for pneumonia hospitalization was 0.90 (95% CI 0.87–0.92) for every 10 mg/dl increase in baseline HDL-C, and 1.02 (95% CI 1.02–1.03) for every 10mg/dl increase in baseline triglycerides. HDL-C and triglycerides both remained significant predictors of pneumonia hospitalization after multivariable adjustment. Such associations were not seen with baseline LDL-C or total cholesterol levels.
Conclusion:
Lower baseline HDL-C and higher triglyceride levels were strongly associated with increased risk of long term pneumonia hospitalization in a large longitudinal US cohort.
Keywords: Lipids, Pneumonia, Observational cohort, ARIC cohort
Introduction
Plasma lipid levels have long been studied in the pathogenesis of atherosclerosis and coronary heart disease (1). However, there is increasing evidence that plasma lipids and lipoproteins also play important roles in host immunity (2–4). Patients with bacterial, viral and parasitic infections have altered plasma lipid levels (5–7), and the degree of alteration in lipid levels correlates with the severity of the infection (8, 9). Studies in patients with active infection demonstrate suppressed high-density lipoprotein (HDL) and total cholesterol concentrations, whereas triglyceride levels are increased (10, 11).
HDL in particular is suggested to be an active participant in the innate immune response (4, 12). Prior observational studies suggest a protective association between HDL cholesterol (HDL-C), and risk of future infection (2–7, 12–20). HDL are immunomodulatory particles containing proteins involved in complement activation and the acute inflammatory response (12, 21). However, studies to date report primarily short-term infectious outcomes associated with HDL levels, mostly within a single hospitalization, and long-term prognostic implications for future infection risk remain unclear. Moreover, there is less information available regarding the relationship of other major lipids to the immune response.
In the current study, we employ a large community-based prospective cohort in the United States including over 13,000 participants followed for a median of >20 years to determine whether baseline major lipid levels are associated with risk of future serious infection measured by hospitalization for pneumonia. The significance of this project lies in the identification of lipid profiles as potential targets for prevention of serious infections in the general population.
Methods
Study population
The Atherosclerosis Risk in Communities (ARIC) study is a large community-based prospective cohort study of 15,792 Americans ages 45 to 64 years recruited from 4 different communities in the US (Forsyth County, NC; Jackson, MS; suburban Minneapolis, MN; and Washington County, MD) (22). The initial examination was conducted between 1987–1989 (visit 1) and subsequent visits took place in 1990–1992 (visit 2), 1993–1995 (visit 3), 1996–1998 (visit 4), and 2011–2013 (visit 5). The study population for this analysis was selected from visit 2 to ensure that participants with a recent prior event of the primary outcome of pneumonia (detailed subsequently) were excluded. However, there was no information regarding events prior to visit 1. Of the 14,348 participants with baseline visit (visit 2) data, participants were excluded for lack of additional follow up information (n=4), incidence of pneumonia prior to visit 2 (n=67), race other than black or white given small numbers (n=41), missing plasma lipid data (HDL-C, LDL-C, total cholesterol and triglyceride levels) (n=332) and non-fasting blood collection (n=426). The final population for the primary analysis consisted of 13,478 participants (Figure 1). The IRB at each participating institution approved the ARIC study and all participants provided informed consent before each examination.
Figure 1.
Flowchart of participants included
Major Lipids
Plasma HDL-C, LDL-C, total cholesterol and triglycerides were measured at a central lipid laboratory at Baylor University. Participants were asked to fast for 12 hours before their visit. Plasma cholesterol and triglyceride were measured by enzymatic methods (23, 24). HDL-C was measured using the Olympus HDL-Cholesterol test (HDL-C), a two-reagent homogenous system that precipitates HDL-C from non-HDL lipoproteins. LDL-C was calculated using Friedewald’s formula (25).
Outcome
The outcome of interest was time from baseline (visit 2) to first incident of hospitalization with pneumonia. Pneumonia events among ARIC participants were identified through hospital discharge records obtained through active surveillance of local hospitals using ICD-9-CM codes 480–486. Participants who were never hospitalized for pneumonia were censored when they were lost to follow up, died or administratively censored at the end of the follow-up period, 12/31/2013.
Other covariates of interest
Covariates of interest were assessed at visit 2 except for center and level of education which was collected at visit 1. Smoking status was determined based on self-reported questionnaires. Use of steroids, antineoplastic agents and cholesterol medications including statins were obtained from medication records at visit 2. Hypertension(HTN) was defined as using an antihypertensive drug, systolic blood pressure ≥140𝑚𝑚𝐻𝑔 or diastolic blood pressure ≥90𝑚𝑚𝐻𝑔. Diabetes(DM) was defined as self-reported physician diagnosis of diabetes, using an antidiabetic drug, fasting glucose level of ≥126mg/dL or random glucose level of ≥200mg/dL. Chronic obstructive pulmonary disease (COPD) was collected from ICD-9-CM codes (490, 491, 492, 494, 496) from hospital discharge records. This information was not available prior to visit 1. eGFR was based on the Chronic Kidney Disease Epidemiology Collaboration serum creatinine and cystatin C equation (26, 27). High-sensitivity C-reactive protein (hs-CRP) was measured at baseline visit using the immunoturbidimetric CRP-Latex(II) hs assay.
Statistical analysis
Baseline characteristics were compared by baseline major lipid quartiles using 𝜒2 tests for categorical and Kruskal Wallis tests for continuous variables. We used Cox proportional hazard analysis to test the association between baseline lipid levels and time to first pneumonia hospitalization. Models were adjusted for demographics, education, smoking, steroids, antineoplastic agents, eGFR and comorbidities (presence of HTN or DM or COPD). Observed incidence of pneumonia hospitalization over time was assessed by baseline lipid quartiles to address potential non-linear effects. Log-rank test was applied to compare survival between quartiles. Sensitivity analysis was performed in participants with baseline hs-CRP values available (N=12,601), and base-10 logarithm transformation was applied when included in models to fit log linearity. Interactions were assessed using two-factor interaction terms between lipid variables that were significantly associated with the primary outcome. Model discrimination was assessed using C-statistic (28). Assumptions were checked and satisfied. All reported p-values are based on two sided tests, and <0.05 was considered statistically significant. Data was processed and analyzed using SAS version 9.4 (SAS institute, Cary, NC).
Results
Pneumonia hospitalization risk and Baseline Characteristics
The median age of participants was 57 years; 56% were female, and 76% were white. Participants were followed for a median of 21.5 (IQR of 15.8–22.6) years, and 1969 (14.6%) participants experienced incident pneumonia hospitalization during follow-up. Comparison of demographics, lifestyle factors, comorbidities between participants by baseline HDL-C (Table 1) and triglyceride quartiles(Table 2) showed that patients in the lower HDL-C quartiles and higher triglyceride quartiles had higher BMI, hs-CRP, higher rates of smoking and comorbidities (HTN, DM, COPD).
Table 1.
Baseline characteristics by HDL quartiles
| Characteristics | HDL quartiles | P-value* | |||
|---|---|---|---|---|---|
| Q1 (<38mg/dL) | Q2 (38–46.9mg/dL) | Q3 (47–58.9mg/dL) | Q4 (≥59mg/dL) | ||
| Age, Median (IQR) | 57(52–62) | 57(52–62) | 56(52–62) | 56(52–61) | <0.01 |
| Race | |||||
| White | 2833(84.2) | 2847(77.8) | 2155(71.5) | 2431(70.7) | <0.01 |
| Gender | |||||
| Female | 912(27.1) | 1720(47.0) | 2042(67.8) | 2819(81.9) | <0.01 |
| Center | |||||
| Forsyth County NC | 956(28.4) | 939(25.7) | 699(23.2) | 864(25.1) | <0.01 |
| Jackson, MS | 463(13.8) | 708(19.4) | 758(25.2) | 894(26.0) | |
| Minneapolis, MN | 861(25.6) | 990(27.1) | 796(26.4) | 997(29.0) | |
| Washington County, MD | 1084(32.2) | 1022(27.9) | 761(25.3) | 686(19.9) | |
| Smoking status | |||||
| Current cigarette smoker | 916(27.2) | 832(22.8) | 599(19.9) | 621(18.1) | <0.01 |
| Former cigarette smoker | 1437(42.8) | 1444(39.5) | 1090(36.2) | 1144(33.3) | |
| Never smoked cigarettes | 1003(29.9) | 1376(37.7) | 1323(43.9) | 1673(48.7) | |
| BMI (kg/m2), Median (IQR) | 28.4(25.8–31.5) | 27.5(24.8–31.0) | 27.0(24.0–30.8) | 25.2(22.5–28.6) | <0.01 |
| Education | |||||
| 12 years or more | 2578(76.8) | 2884(78.9) | 2341(77.8) | 2794(81.3) | <0.01 |
| Medications | |||||
| Cholesterol lowering medication including Statins | 271(8.1) | 261(7.2) | 178(5.9) | 152(4.4) | <0.01 |
| Steroids | 19(0.6) | 46(1.3) | 38(1.3) | 79(2.3) | <0.01 |
| Antineoplastic agents | 17(0.5) | 22(0.6) | 14(0.5) | 25(0.7) | 0.51 |
| Comorbidities | |||||
| Hypertension | 1334(39.8) | 1325(36.3) | 1042(34.7) | 1045(30.5) | <0.01 |
| Diabetes | 681(20.3) | 564(15.5) | 358(11.9) | 216(6.3) | <0.01 |
| COPD | 3(0.1) | 4(0.1) | 0(0) | 3(0.1) | 0.4 |
| Patients with comorbiditiesτ | 1633(48.7) | 1564(42.9) | 1182(39.3) | 1139(33.3) | <0.01 |
| eGFR (mL/min/1.73m2), Median (IQR) | 95.5(85.5–102.9) | 96.6(88.0–104.6) | 98.5(89.8–107.4) | 99.6(91.4–107.8) | <0.01 |
| Hs-CRP (mg/dL) )ξ, Median (IQR) | 2.52(1.28–5.06) | 2.27(1.13–4.91) | 2.22(1.04–4.88) | 1.80(0.83–4.18) | <0.01 |
Chi-square tests were used for categorical variables and Wilcoxon rank-sum tests were applied for continuous variables.
hsCRP collected in total N=12,601 (pneumonia event yes N=1805, no N=10796)
Patients with comorbidities include patients with at least 1 comorbid condition
Table 2.
Baseline characteristics by Triglyceride quartiles
| Characteristics | Triglyceride quartiles | P-value* | |||
|---|---|---|---|---|---|
| Q1 (<80.2mg/dL) | Q2 (80.2–110.8mg/dL) | Q3 (110.9–157.4mg/dL) | Q4 ( 157.5mg/dL) | ||
| Age, Median (IQR) | 56(51–61) | 57(52–62) | 57(53–62) | 57(53–62) | <0.01 |
| Race | |||||
| White | 2250(66.3) | 2461(73.4) | 2673(78.8) | 2882(86.3) | <0.01 |
| Gender | |||||
| Female | 2012(59.3) | 1940(57.9) | 1875(55.3) | 1666(49.9) | <0.01 |
| Center | |||||
| Forsyth County NC | 789(23.2) | 867(25.9) | 896(26.4) | 906(27.1) | <0.01 |
| Jackson, MS | 1003(29.5) | 778(23.2) | 632(18.6) | 410(12.3) | |
| Minneapolis, MN | 915(27.0) | 884(26.4) | 915(27.0) | 930(27.8) | |
| Washington County, MD | 688(20.3) | 822(24.5) | 949(28.0) | 1094(32.8) | |
| Smoking status | |||||
| Current cigarette smoker | 673(19.9) | 751(22.5) | 777(22.9) | 767(23.0) | <0.01 |
| Former cigarette smoker | 1215(35.8) | 1242(37.1) | 1287(38.0) | 1371(41.1) | |
| Never smoked cigarettes | 1503(44.3) | 1351(40.4) | 1325(39.1) | 1196(35.9) | |
| BMI (kg/m2), Median (IQR) | 25.5(22.9–28.7) | 26.7(23.8–30.2) | 27.5(24.9–31.0) | 28.6(25.8–31.9) | <0.01 |
| Education | |||||
| 12 years or more | 2719(80.2) | 2589(77.4) | 2658(78.6) | 2631(78.8) | 0.05 |
| Medications | |||||
| Cholesterol lowering medication including Statins | 139(4.1) | 158(4.7) | 231(6.8) | 334(10.0) | <0.01 |
| Steroids | 43(1.3) | 51(1.5) | 43(1.3) | 45(1.4) | 0.78 |
| Antineoplastic agents | 17(0.5) | 17(0.5) | 22(0.7) | 22(0.7) | 0.73 |
| Comorbidities | |||||
| Hypertension | 950(28.1) | 1089(32.6) | 1243(36.7) | 1464(44.0) | <0.01 |
| Diabetes | 229(6.8) | 333(10.0) | 480(14.2) | 777(23.3) | <0.01 |
| COPD | 1(0.03) | 4(0.12) | 2(0.06) | 3(0.09) | 0.56 |
| Patients with comorbiditiesτ | 1055(31.2) | 1239(37.2) | 1455(43.0) | 1769(53.2) | <0.01 |
| eGFR (mL/min/1.73m2), Median (IQR) | 99.8(91.4–108.9) | 97.8(88.8–105.8) | 96.3(87.3–104.3) | 95.8(86.6–103.3) | <0.01 |
| Hs-CRP (mg/dL) )ξ, Median (IQR) | 1.54(0.74–3.51) | 2.04(0.98–4.35) | 2.51(1.22–5.21) | 2.82(1.45–5.75) | <0.01 |
Chi-square tests were used for categorical variables and Wilcoxon rank-sum tests were applied for continuous variables.
hsCRP collected in total N=12,601 (pneumonia event yes N=1805, no N=10796)
Patients with comorbidities include patients with at least 1 comorbid condition
Pneumonia hospitalization risk and Baseline Lipid levels
Participants who had pneumonia hospitalizations had lower baseline HDL-C levels and higher triglyceride levels compared with those who did not have pneumonia events (Table S1). When divided into quartiles, subjects with the lowest baseline HDL-C (HDL-C 1st quartile, Q1) had significantly more pneumonia hospitalization events(n=572, 17%) compared to subjects with higher baseline HDL-C levels (Q2: 565(15%); Q3: 415(14%); Q4: 417(12%), p<0.01, Figure 2–a). For triglyceride quartiles, there was a graded increased risk of pneumonia the higher the baseline triglyceride quartile (Q1: 432(13%); Q2: 461(14%); Q3: 503(15%); Q4: 573(17%), p<0.01, Figure 2–d). Such trends were not seen in the LDL-C quartiles or the total cholesterol quartiles (Figure 2–b, c).
Figure 2.
Pneumonia hospitalization cumulative hazard by major lipid quartiles
In Cox proportional hazard models, hazard ratio (HR) of pneumonia hospitalization was 0.90 (95% CI 0.87–0.92) for every 10 mg/dl increase in baseline HDL-C. Triglyceride also showed a significant association with pneumonia hospitalization, with HR of 1.02 (95% CI 1.02–1.03) for every 10 mg/dl increase in baseline triglyceride levels (Table 3). LDL-C and total cholesterol were not associated with risk of pneumonia hospitalizations.
Table 3.
Hazard ratios of pneumonia hospitalization
| Covariate set | HR (95% CI) | |
|---|---|---|
|
HDL-C
(Unit: 10 mg/dL) |
Unadjusted | 0.90 (0.87, 0.92)** |
| Covariate set 1 | 0.93 (0.90, 0.96)** | |
| Covariate set 2 | 0.97 (0.94, 1.00)* | |
|
LDL-C
(Unit: 10 mg/dL) |
Unadjusted | 1.01 (1.00, 1.02) |
| Covariate set 1 | 1.00 (0.99, 1.01) | |
| Covariate set 2 | 1.00 (0.99, 1.01) | |
|
Total cholesterol
(Unit: 10 mg/dL) |
Unadjusted | 1.00 (0.99, 1.02) |
| Covariate set 1 | 1.00 (0.99, 1.01) | |
| Covariate set 2 | 1.00 (0.99, 1.01) | |
|
Triglyceride
(Unit: 10 mg/dL) |
Unadjusted | 1.02 (1.02,1.03)** |
| Covariate set 1 | 1.02 (1.01, 1.03)** | |
| Covariate set 2 | 1.01 (1.00, 1.02)* | |
p<0.05
p<0.001
Reported HR is per 10mg/dL change in lipid level.
Covariate set 1: adjusted for demographics including age, race, gender, center
Covariate set 2: adjusted for Covariate set 1 and education, smoking, steroid, antineoplastic agents, eGFR, comorbidities (presence of diabetes or hypertension or COPD)
In multivariable analysis adjusted for demographics (age, gender, race, center), smoking, education level, medications, renal function and comorbidities, baseline low HDL-C and high triglyceride remained a significant predictor of the risk of pneumonia hospitalization (Table 3, Supplementary Table 2). The graded association between pneumonia with HDL-C and triglyceride was observed across the entire range of HDL-C and triglycerides (Supplementary Figure).
Elevated levels of hs-CRP is a widely used biomarker of acute inflammation, and has been demonstrated to be independently associated with higher risk of future infection (29). In a sensitivity analysis restricting to participants with baseline hs-CRP values available (N=12,601), triglycerides maintained significant HRs in all models (data not shown).
Pneumonia hospitalization risk according to HDL-C by triglyceride quartiles
As HDL-C and triglycerides both remained strong predictors of the primary outcome, we tested whether there was an interaction between HDL and triglyceride and found it to be non-significant. Nonetheless, we examined the risk of pneumonia hospitalizations according to HDL-C quartiles stratified by each triglyceride quartile, as the function of HDL particles can be altered by their triglyceride content. The association between HDL-C concentration and risk of pneumonia hospitalizations remained strong regardless of triglyceride quartiles (Figure 3). Participants in HDL-C Q4 had the lowest incidence of pneumonia hospitalizations within each triglyceride quartile. In triglyceride quartile 2 and 3, incidence of pneumonia hospitalization increased sequentially from HDL Q4 to Q1. Within triglyceride Q1-Q3, subjects in lower HDL-C quartiles trended towards higher adjusted HR for pneumonia. Interestingly, most participants in the highest triglyceride quartile (triglyceride Q4) had significantly increased risk of pneumonia hospitalizations regardless of HDL-C.
Figure 3.
Risk of pneumonia events according to HDL by triglyceride quartiles
Referent group is triglyceride Quartile 1-HDL Quartile 4
Adjusted for covariate set 1: age, gender, race, center
Discussion
In the current study, we demonstrate that differences in certain baseline lipid levels are associated with a higher risk of future pneumonia hospitalizations in a large prospective cohort of 4 US communities. Participants with low HDL-C or high triglyceride levels at baseline had a significantly increased risk of future pneumonia hospitalizations over a median follow up time of >20 years, independent of other known risk factors such as age, smoking status, antineoplastic medication use, steroid use, renal dysfunction or presence of comorbidities. Such associations were not seen for LDL-C or total cholesterol.
Our results support the hypothesis that there is a link between lipid metabolism and the immune system. Studies have demonstrated decreased lipoprotein cholesterol levels in patients with critical illness, severe sepsis and other acute infections. (5, 6, 10, 11, 30, 31). HDL in particular has been suggested to play an important role as a part of the innate immune response using rapid induction of an oxidative state as a way of combating pathogens (32, 33). HDL also demonstrates anti-inflammatory, antioxidant properties by affecting the expression of local adhesion molecules and cytokine secretion by immune cells (33, 34). Decreased HDL-C, has been shown to have prognostic implications in short term studies of hospitalized patients. Patients with lower baseline HDL-C on admission had increased rates of infectious complications, intensive care unit (ICU) admissions and mortality during hospitalization (6, 15, 31, 33, 35–37). An inverse correlation between HDL-C levels at admission and plasma acute phase reactants such as CRP has been reported in a retrospective study of 107 patients hospitalized with community-acquired pneumonia at a University hospital in Spain (37), and was also observed in our current study. In the former retrospective study, low HDL baseline levels at hospital admission associated with development of septic shock, ICU admission, and the presence of a pleural effusion complicating the pneumonia.
Fewer studies have looked at the association between serum lipids and long-term risk of developing infections. A recent study of two large Danish cohorts including 97,166 individuals from the Copenhagen General Population(CGP) Study and 9,387 individuals from the Copenhagen City Heart(CCH) Study demonstrated a U-shaped association between HDL-C levels and risk of hospitalization from infection, over a median follow up times of 6 and 20 years respectively (20). In the CGP cohort, the HR for any infection was 1.75 (95% CI 1.31–2.34) in subjects with the lowest HDL-C (<31 mg/dL) and 1.43 (1.16–1.76) in subjects with the highest HDL-C (≥100 mg/dL) compared to the referent group (HDL-C=85–95 mg/dL). Among different types of infections, gastroenteritis and bacterial pneumonia had the highest risk estimates. Similar associations between low and high HDL-C levels and risk of infectious disease were confirmed in the CCH cohort. This data supports the hypothesis that circulating lipids may play an important long-term role in the body’s defense against infectious pathogens. However, certain points should be considered in the interpretation of these findings. Participants in the Danish study were all white individuals of Danish descent, whereas the ARIC cohort had a broader ethnic population including 24% blacks. Also, participants in ARIC had lower median HDL-C (47mg/dL in ARIC vs. 54–62 mg/dL in the CGP cohort), higher proportions of obesity (BMI ≥30; 28.4% in ARIC vs. 16% in the CGP cohort) and diabetes (13.5% in ARIC vs 4% in the CGP cohort) which is reflective of the higher prevalence of metabolic syndrome in America compared to Europe. Also, ascertainment of the primary outcome in ARIC was through active surveillance of the cohort rather than from a separate registry as was done in the Danish study, which further increases its reliability. Our study did not demonstrate the U-shaped association between HDL-C levels and risk of infections which may partially be explained by the small number of patients in ARIC with extremely high HDL-C levels (above 100 mg/dL). Nonetheless, our findings show HDL-C levels are strongly associated with long-term risk of serious infections while LDL-C and total cholesterol levels are not, which is in agreement with the Danish study.
The current work is the first large study in the general population to demonstrate a strong, significant association between higher baseline triglyceride levels and risk of future infection. Patients with elevated triglycerides are at increased risk for cardiovascular mortality (38) and lowering triglycerides leads to a significant reduction in ischemic events (39). However, the effect of triglyceride levels in regard to inflammatory states is less consistent between studies. While some studies show increased triglyceride levels during acute infections (10, 11), other studies including the Danish study have demonstrated no significant association between triglyceride levels and infectious outcomes (6, 20). Most studies are limited in sample size and follow up time. In the Danish study, triglyceride levels were not associated with infectious risks after multivariate adjustment. Our study demonstrated that higher triglycerides remained strongly associated with increased risk of hospitalization for pneumonia after multivariate adjustment. The current study is also strengthened by our ability to rigorously measure the lipid profiles in fasting condition excluding the possibility of postprandial effects on triglycerides, which was not done in other studies.
Studies have previously shown that the composition of HDL-C is modified to have higher triglyceride content in patients with hypertriglyceridemia as well as in acute inflammatory states (40). Such triglyceride-enriched HDL-C particles appear to be aberrant in their functional properties including their anti-oxidative function (41). Here we evaluated the risk of pneumonia according to HDL-C quartiles within each triglyceride quartile and found that high HDL-C was consistently associated with lower risk of pneumonia in all triglyceride quartiles. This supports the a priori hypothesis that HDL-C plays a protective role in the host immune response. Interestingly, participants in the highest triglyceride quartile had significantly increased risk of pneumonia regardless of HDL-C quartile. We postulate that subjects in the highest triglyceride quartile may represent subjects with altered HDL function due to triglyceride enrichment which may be more influential than the HDL level itself. Further studies are needed to gain a better understanding of how triglyceride enriched lipoproteins may modulate the host immune response.
While current lipid modifying therapies are largely used to prevent cardiovascular events, the current findings suggest they may also generate additional benefit in preventing future infections. Although our findings cannot confirm that initiating agents to modify HDL-C or triglyceride levels lower the risk of future infections, they encourage maintaining healthy lipid profiles for reasons beyond cardiovascular prevention. In addition, when prescribing non-lipid medications that modify lipid levels physicians may consider medications that increase HDL-C if possible, particularly in patients with high risk of future infections. Absolute benefits of increasing HDL-C are likely to be greater in patients without extremely high triglyceride levels.
Our study has several limitations. First, although we excluded patients with known pneumonia hospitalizations, initial cohort enrollment was not designed to assess infectious outcomes and we do not have information about pneumonia events before the start of the ARIC study; thus the association between baseline lipid levels and pneumonia hospitalizations cannot be established as causal. Second, use of a population-based cohort allowed us to study a large number of participants over a long follow up period, but there were constraints in age and ethnicity. For generalizability of our findings, future studies should be considered in more diverse populations. Lastly, we censored patients after their first episode of pneumonia hospitalization; assessment of repeated events or competing risks of other illnesses/death may have an effect on risk estimates.
In conclusion, after analyzing 13,478 participants in a longitudinal prospective US cohort, we found that lower HDL-C and higher triglyceride levels at baseline are associated with a significantly increased risk of future pneumonia. Our study provides evidence that alterations in certain lipids may play important roles in modulating the host’s immune response. Future studies are needed to assess the efficacy of improving serum lipid levels by lifestyle and pharmacologic interventions in modifying future infection risk.
Supplementary Material
Highlights.
Lipids are implicated as important modulators of immune response
Longitudinal observation cohort study to asses risk of pneumonia
Lower HDL-C and higher triglyceride associated with higher risk of future pneumonia
Altered lipids have long term implications in future infections
Acknowledgement
This project was supported by the UCLA Older American Independence Center (NIH P30 AG028748) and NIH National Center for Advancing Translational Science (NCATS) UCLA CTSI Grant Number UL1TR001881. The Atherosclerosis Risk in Communities study has been funded in whole or in part with Federal funds from the NHLBI, NIH, Department of Health and Human Services (HHSN268201700001I, HHSN268201700002I, HHSN268201700003I, HHSN268201700005I, HHSN268201700004I). The authors thank the staff and participants of the ARIC study for their important contributions.
Funding source: Dr. Charles-Schoeman received support from the NHLBI (R01HL123064). This project was supported in part by the UCLA Older American Independence Center (NIH P30 AG028748) and NIH National Center for Advancing Translational Science (NCATS) UCLA CTSI Grant Number UL1TR001881. The Atherosclerosis Risk in Communities study has been funded in whole or in part with Federal funds from the NHLBI, NIH, Department of Health and Human Services (HHSN268201700001I, HHSN268201700002I, HHSN268201700003I, HHSN268201700005I, HHSN268201700004I). Funding source had no participation in study design; in the collection, analysis, and interpretation of data; in writing the report; or in the decision to submit the article for publication.
Footnotes
There was no conflict of interest for all authors.
All authors had access to the data and participated in preparation of the manuscript.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Reference
- 1.Yusuf S, Hawken S, Ounpuu S, Dans T, Avezum A, Lanas F, et al. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study. Lancet. 2004;364(9438):937–52. [DOI] [PubMed] [Google Scholar]
- 2.Khovidhunkit W, Kim MS, Memon RA, Shigenaga JK, Moser AH, Feingold KR, et al. Effects of infection and inflammation on lipid and lipoprotein metabolism: mechanisms and consequences to the host. J Lipid Res. 2004;45(7):1169–96. [DOI] [PubMed] [Google Scholar]
- 3.Han R. Plasma lipoproteins are important components of the immune system. Microbiol Immunol. 2010;54(4):246–53. [DOI] [PubMed] [Google Scholar]
- 4.Murphy AJ, Chin-Dusting JPF, Sviridov D, Woollard KJ. The anti inflammatory effects of high density lipoproteins,. Current Medicinal Chemistry. 2009;16(6):667–75. [DOI] [PubMed] [Google Scholar]
- 5.Gallin JI, Kaye D, O’Leary WM. Serum lipids in infection. N Engl J Med. 1969;281(20):1081–6. [DOI] [PubMed] [Google Scholar]
- 6.Gordon BR, Parker TS, Levine DM, Saal SD, Wang JC, Sloan BJ, et al. Low lipid concentrations in critical illness: implications for preventing and treating endotoxemia. Crit Care Med. 1996;24(4):584–9. [DOI] [PubMed] [Google Scholar]
- 7.van Leeuwen HJ, Heezius EC, Dallinga GM, van Strijp JA, Verhoef J, van Kessel KP. Lipoprotein metabolism in patients with severe sepsis. Crit Care Med. 2003;31(5):1359–66. [DOI] [PubMed] [Google Scholar]
- 8.Deniz O, Tozkoparan E, Yaman H, Cakir E, Gumus S, Ozcan O, et al. Serum HDL-C levels, log (TG/HDL-C) values and serum total cholesterol/HDL-C ratios significantly correlate with radiological extent of disease in patients with community-acquired pneumonia. Clin Biochem. 2006;39(3):287–92. [DOI] [PubMed] [Google Scholar]
- 9.Deniz O, Gumus S, Yaman H, Ciftci F, Ors F, Cakir E, et al. Serum total cholesterol, HDL-C and LDL-C concentrations significantly correlate with the radiological extent of disease and the degree of smear positivity in patients with pulmonary tuberculosis. Clin Biochem. 2007;40(3–4):162–6. [DOI] [PubMed] [Google Scholar]
- 10.Rodriguez Reguero JJ, Iglesias Cubero G, Vazquez M, Folgueras I, Braga S, Bustillo E, et al. Variation in plasma lipid and lipoprotein concentrations in community-acquired pneumonia a six-month prospective study. Eur J Clin Chem Clin Biochem. 1996;34(3):245–9. [DOI] [PubMed] [Google Scholar]
- 11.Sammalkorpi K, Valtonen V, Kerttula Y, Nikkila E, Taskinen MR. Changes in serum lipoprotein pattern induced by acute infections. Metabolism. 1988;37(9):859–65. [DOI] [PubMed] [Google Scholar]
- 12.Norata GD, Pirillo A, Ammirati E, Catapano AL. Emerging role of high density lipoproteins as a player in the immune system. Atherosclerosis. 2012;220(1):11–21. [DOI] [PubMed] [Google Scholar]
- 13.Barcia AM, Harris HW. Triglyceride-rich lipoproteins as agents of innate immunity. Clin Infect Dis. 2005;41 Suppl 7:S498–503. [DOI] [PubMed] [Google Scholar]
- 14.Barlage S, Gnewuch C, Liebisch G, Wolf Z, Audebert FX, Gluck T, et al. Changes in HDL-associated apolipoproteins relate to mortality in human sepsis and correlate to monocyte and platelet activation. Intensive Care Med. 2009;35(11):1877–85. [DOI] [PubMed] [Google Scholar]
- 15.Chien JY, Jerng JS, Yu CJ, Yang PC. Low serum level of high-density lipoprotein cholesterol is a poor prognostic factor for severe sepsis. Crit Care Med. 2005;33(8):1688–93. [DOI] [PubMed] [Google Scholar]
- 16.Grion CM, Cardoso LT, Perazolo TF, Garcia AS, Barbosa DS, Morimoto HK, et al. Lipoproteins and CETP levels as risk factors for severe sepsis in hospitalized patients. Eur J Clin Invest. 2010;40(4):330–8. [DOI] [PubMed] [Google Scholar]
- 17.Gruber M, Christ-Crain M, Stolz D, Keller U, Muller C, Bingisser R, et al. Prognostic impact of plasma lipids in patients with lower respiratory tract infections - an observational study. Swiss Med Wkly. 2009;139(11–12):166–72. [DOI] [PubMed] [Google Scholar]
- 18.Rodriguez-Sanz A, Fuentes B, Martinez-Sanchez P, Prefasi D, Martinez-Martinez M, Correas E, et al. High-density lipoprotein: a novel marker for risk of in-hospital infection in acute ischemic stroke patients? Cerebrovasc Dis. 2013;35(3):291–7. [DOI] [PubMed] [Google Scholar]
- 19.Shor R, Wainstein J, Oz D, Boaz M, Matas Z, Fux A, et al. Low HDL levels and the risk of death, sepsis and malignancy. Clin Res Cardiol. 2008;97(4):227–33. [DOI] [PubMed] [Google Scholar]
- 20.Madsen CM, Varbo A, Tybjaerg-Hansen A, Frikke-Schmidt R, Nordestgaard BG. U-shaped relationship of HDL and risk of infectious disease: two prospective population-based cohort studies. Eur Heart J. 2018;39(14):1181–90. [DOI] [PubMed] [Google Scholar]
- 21.Vaisar T, Pennathur S, Green PS, Gharib SA, Hoofnagle AN, Cheung MC, et al. Shotgun proteomics implicates protease inhibition and complement activation in the antiinflammatory properties of HDL. J Clin Invest. 2007;117(3):746–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.The ARIC Investigators. The Atherosclerosis Risk in Communities (ARIC) Study: design and objectives. Am J Epidemiol. 1989;129(April):687–702. [PubMed] [Google Scholar]
- 23.Allain CC, Poon LS, Chan CS, Richmond W, Fu PC. Enzymatic determination of total serum cholesterol. Clin Chem. 1974;20(4):470–5. [PubMed] [Google Scholar]
- 24.Bucolo G, David H. Quantitative determination of serum triglycerides by the use of enzymes. Clin Chem. 1973;19(5):476–82. [PubMed] [Google Scholar]
- 25.Tietz NW. Textbook of Clinical Chemistry, WB Saunders. 1986:888. [Google Scholar]
- 26.Inker LA, Schmid CH, Tighiouart H, Eckfeldt JH, Feldman HI, Greene T, et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med. 2012;367(1):20–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Ishigami J, Grams ME, Chang AR, Carrero JJ, Coresh J, Matsushita K. CKD and Risk for Hospitalization With Infection: The Atherosclerosis Risk in Communities (ARIC) Study. Am J Kidney Dis. 2017;69(6):752–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Liu L, Forman S, Barton B. Fitting Cox Model Using PROC PHREG and Beyond in SAS. SAS Global Forum 2009. 2009(236). [Google Scholar]
- 29.Wang HE, Shapiro NI, Safford MM, Griffin R, Judd S, Rodgers JB, et al. High-sensitivity C-reactive protein and risk of sepsis. PLoS One. 2013;8(7):e69232. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Chien YF, Chen CY, Hsu CL, Chen KY, Yu CJ. Decreased serum level of lipoprotein cholesterol is a poor prognostic factor for patients with severe community-acquired pneumonia that required intensive care unit admission. J Crit Care. 2015;30(3):506–10. [DOI] [PubMed] [Google Scholar]
- 31.Giovannini I, Boldrini G, Chiarla C, Giuliante F, Vellone M, Nuzzo G. Pathophysiologic correlates of hypocholesterolemia in critically ill surgical patients. Intensive Care Med. 1999;25(7):748–51. [DOI] [PubMed] [Google Scholar]
- 32.Navab M, Anantharamaiah GM, Fogelman AM. The role of high-density lipoprotein in inflammation. Trends Cardiovasc Med. 2005;15(4):158–61. [DOI] [PubMed] [Google Scholar]
- 33.Catapano AL, Pirillo A, Bonacina F, Norata GD. HDL in innate and adaptive immunity. Cardiovasc Res. 2014;103(3):372–83. [DOI] [PubMed] [Google Scholar]
- 34.Sviridov D, Remaley AT. High-density lipoprotein mimetics: promises and challenges. Biochem J. 2015;472(3):249–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Gui D, Spada PL, De Gaetano A, Pacelli F. Hypocholesterolemia and risk of death in the critically ill surgical patient. Intensive Care Med. 1996;22(8):790–4. [DOI] [PubMed] [Google Scholar]
- 36.Peng YS, Chen YC, Tian YC, Yang CW, Lien JM, Fang JT, et al. Serum levels of apolipoprotein A-I and high-density lipoprotein can predict organ failure in acute pancreatitis. Crit Care. 2015;19:88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Saballs M, Parra S, Sahun P, Pelleja J, Feliu M, Vasco C, et al. HDL-c levels predict the presence of pleural effusion and the clinical outcome of community-acquired pneumonia. Springerplus. 2016;5(1):1491. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Klempfner R, Erez A, Sagit BZ, Goldenberg I, Fisman E, Kopel E, et al. Elevated Triglyceride Level Is Independently Associated With Increased All-Cause Mortality in Patients With Established Coronary Heart Disease: Twenty-Two-Year Follow-Up of the Bezafibrate Infarction Prevention Study and Registry. Circ Cardiovasc Qual Outcomes. 2016;9(2):100–8. [DOI] [PubMed] [Google Scholar]
- 39.Bhatt DL, Steg PG, Miller M, Brinton EA, Jacobson TA, Ketchum SB, et al. Cardiovascular Risk Reduction with Icosapent Ethyl for Hypertriglyceridemia. N Engl J Med. 2019;380(1):11–22. [DOI] [PubMed] [Google Scholar]
- 40.Skeggs JW, Morton RE. LDL and HDL enriched in triglyceride promote abnormal cholesterol transport. J Lipid Res. 2002;43(8):1264–74. [PubMed] [Google Scholar]
- 41.Imaizumi S, Navab M, Morgantini C, Charles-Schoeman C, Su F, Gao F, et al. Dysfunctional high-density lipoprotein and the potential of apolipoprotein A-1 mimetic peptides to normalize the composition and function of lipoproteins. Circ J. 2011;75(7):1533–8. [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.