Abstract
Persons with chronic kidney disease (CKD) are at high risk of infection. While low-grade inflammation could impair immune response, it is unknown whether inflammatory markers are associated with infection risk in this clinical population. Using 2003–2013 data from the Chronic Renal Insufficiency Cohort Study (3,597 participants with CKD), we assessed the association of baseline plasma levels of 4 inflammatory markers (interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α), interleukin-1 receptor antagonist (IL-1RA), and transforming growth factor-β (TGF-β)) with incident hospitalization with major infection (pneumonia, urinary tract infection, cellulitis and osteomyelitis, and bacteremia and sepsis). During follow-up (median 7.5 years), 36% (n = 1,290) had incident hospitalization with major infection. In multivariable Cox analyses with each inflammatory marker modeled as a restricted cubic spline, higher levels of IL-6 and TNF-α were monotonically associated with increased risk of hospitalization with major infection (for 95th vs. 5th percentile, hazard ratio = 2.11 (95% confidence interval: 1.68, 2.66) for IL-6 and 1.88 (95% confidence interval: 1.51, 2.33) for TNF-α), while corresponding associations for IL-1RA or TGF-β were nonsignificant. Thus, higher plasma levels of IL-6 and TNF-α, but not IL-1RA or TGF-β, were significantly associated with increased risk of hospitalization with major infection. Future studies should investigate whether inflammatory pathways that involve IL-6 and TNF-α increase susceptibility to infection among individuals with CKD.
Keywords: chronic kidney disease, chronic renal insufficiency, infection, infectious disease, interleukin-1 receptor antagonist, interleukin-6, transforming growth factor-β, tumor necrosis factor-α
Abbreviations
- CI
confidence interval
- CKD
chronic kidney disease
- CRIC
Chronic Renal Insufficiency Cohort
- eGFR
estimated glomerular filtration rate
- ICD-9-CM
International Classification of Diseases, Ninth Revision, Clinical Modification
- IL-1RA
interleukin-1 receptor antagonist
- IL-6
interleukin-6
- TGF-β
transforming growth factor-β
- TNF-α
tumor necrosis factor-α
- uACR
urinary albumin-to-creatinine ratio
Infectious disease poses a major public health burden on persons with chronic kidney disease (CKD). Among patients with end-stage renal disease, infection is the most frequent cause of hospitalization, and the second leading cause of death (1). In addition, we recently demonstrated that mild to moderate CKD was associated with hospitalization with infection and subsequent infection-related mortality (2, 3). While uremic toxins and vascular access might explain most of the increased risk of infection in end-stage renal disease (4), underlying mechanisms of increased risk of infection in less severe CKD are not well understood (5).
CKD is often a low-grade inflammatory state with elevated levels of inflammatory markers such as interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α), interleukin-1 receptor antagonist (IL-1RA), and transforming growth factor-β (TGF-β) (6, 7), which might increase susceptibility to infection in this population. IL-6 and TNF-α are proinflammatory cytokines that are essential for the host defense against pathogens (8, 9); however, unregulated exposure to these cytokines can also impair the immune response (10–12). IL-1RA and TGF-β act as antiinflammatory markers and could interfere with the immune system through the modulation of T-cell regulation (13, 14). A few studies have investigated the association of IL-6 or TNF-α with risk of pneumonia and sepsis, but they were limited by small sample size (n < 400), selected study population of human immunodeficiency virus patients, and case-control design (15–17). Whether these inflammatory markers are associated with risk of infection in persons with CKD is unknown.
We hypothesized that inflammatory markers are each associated with the risk of infection independently of major confounders including estimated glomerular filtration rate (eGFR) and urine albumin-to-creatinine ratio (uACR) among individuals with CKD who are not dialysis-dependent. To test this hypothesis, we conducted a prospective analysis to examine the association of blood levels of 4 inflammatory markers (IL-6, TNF-α, IL-1RA, and TGF-β) with risk of incident hospitalization with major infection (pneumonia, urinary tract infections, cellulitis and osteomyelitis, and bacteremia and sepsis) among individuals in the Chronic Renal Insufficiency Cohort (CRIC) Study.
METHODS
The CRIC Study is a prospective cohort established to examine the risk factors for progression of CKD and cardiovascular disease in patients with CKD (18). In brief, between May 2003 and March 2007, CRIC enrolled 3,939 individuals aged 21–74 years with CKD, not dialysis-dependent, and with an eGFR of 20–70 mL/min/1.73 m2 from the 7 study sites across the United States. Recruitment strategies included electronic health-care record searches, manual reviews of medical charts, and referrals from health-care providers. Thus, the study population consisted of those with newly diagnosed CKD as well as prevalent CKD. Of 3,939 participants, we excluded persons with missing inflammatory marker levels (n = 81) and covariates (n = 261). After the exclusions, 3,597 participants were included in the analysis (Web Figure 1, available at https://academic.oup.com/aje). Characteristics were overall comparable between those with and without missing values (Web Table 1). The study was conducted in compliance with the Declaration of Helsinki. Written informed consent for CRIC was obtained from all participants, and the institutional review boards at all study sites approved the study.
Exposures: inflammatory markers
Primary exposures were plasma levels of IL-6, TNF-α, IL-1RA, and TGF-β. Using plasma samples collected at a baseline visit and stored at −80°C according to the standard protocol, high-sensitivity enzyme-linked immunosorbent assays (ELISAs) were used to measure levels of IL-6 and TNF-α and standard sandwich ELISAs were used for IL-1RA and TGF-β (Quantikine HS; R&D Systems, Minneapolis, Minnesota). The lowest detectable levels were 0.07 pg/mL for IL-6, 0.11 pg/mL for TNF-α, 6.3 pg/mL for IL-1RA, and 4.6 pg/mL for TGF-β. All assays were performed in duplicate samples, and the intraassay coefficients of variation were <13% for IL-6, 15.2% for TNF-α, <13% for IL-1RA, and 21.5% for TGF-β.
Outcomes
The primary outcome was the first incidence of hospitalization with major infection defined by the 4 most common types of infection (19) (pneumonia, urinary tract infections, cellulitis and osteomyelitis, and bacteremia and sepsis). Hospitalizations with infection were ascertained using codes from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM), on discharge records: pneumonia (ICD-9-CM codes 480–486), urinary tract infections (ICD-9-CM codes 590.1, 299.0, and 601.0), cellulitis and osteomyelitis (ICD-9-CM codes 040.0, 681, 682, 730.0, and 730.2), and bacteremia and sepsis (ICD-9-CM codes 038, 003.1, 020.2, 022.3, 036.2, 054.5, 785.52, 790.7, 995.91, and 995.92). For primary analysis, we included all infection diagnoses on discharge records regardless of their diagnostic position. Follow-up was started with the date of the baseline visit and continued until December 31, 2013. Those who did not have the primary outcome were censored when they died, were lost-to-follow-up, or at the end of follow-up.
Covariates
All covariates were assessed at the study baseline. Age, sex, race, body mass index, smoking status (never or former smokers vs. current smokers), alcohol consumption, hypertension, and diabetes; past medical history of chronic obstructive pulmonary disease, cancer, cardiovascular disease (heart failure, myocardial infarction, coronary revascularization, and stroke), and steroid use; and cause of CKD, years since CKD diagnosis, eGFR, and uACR were recorded at baseline (20). Hypertension was defined by systolic blood pressure ≥140 mm Hg, diastolic blood pressure ≥90 mm Hg, or self-report of taking prescribed hypertension medication. eGFR was calculated using the CKD Epidemiology Collaboration equation (21). Serum creatinine was measured by the Jaffe method on a Beckman Synchron System (Beckman Coulter, Brea, California). Serum cystatin C was measured on a Dade-Behring BNII (Siemens Healthcare Diagnostic, Tarrytown, New York). Random spot uACR was calculated by dividing urine albumin by urine creatinine.
Statistical analysis
Incidence rates and 95% confidence intervals were estimated using Poisson regression models. Hazard ratios were estimated using Cox proportional hazards models stratifying by center to allow for center-specific baseline hazard. Model 1 adjusted for age, sex, and race. Model 2 additionally accounted for body mass index, smoking status, alcohol consumption, hypertension, diabetes, chronic obstructive pulmonary disease, cancer, cardiovascular disease, corticosteroid use, eGFR, and uACR. To account for possible nonlinear confounding, we modeled eGFR and uACR as linear spline terms with a knot at 60 mL/min/1.73 m2 for eGFR and 30 mg/g for uACR. Model 3 further adjusted for the remaining inflammatory markers, such as TNF-α, IL-1RA, and TGF-β for analysis of IL-6. The levels of inflammatory markers were grouped into 4 categories according to their quartiles, and the lowest quartile served as the referent. We also considered the inflammatory markers as continuous variables, for which they were log-transformed and modeled as a restricted cubic spline with knots at values corresponding to the 5th, 50th, and 95th percentiles. Because there are no established reference ranges for these markers, the referent was set at a value corresponding to the 5th percentile.
To explore the robustness of our primary findings, we performed several sensitivity analyses. First, to more explicitly evaluate the associations of inflammatory markers with hospitalization due to infection, we restricted our analysis to infections as the principal diagnosis. Second, to evaluate the impact of interim cardiovascular events and end-stage renal disease during follow-up on the associations of inflammatory markers with infection risk, we censored individuals with those interim events because they might have influenced the association through different mechanisms (22). Incident end-stage renal disease was defined as kidney transplant or dialysis. Third, to minimize the chance of reverse causation (elevated inflammatory marker levels in those with latent, undiagnosed, or recent infection), we excluded events that occurred within 1 or within 3 years after the baseline visit. Fourth, we limited follow-up to the first 1 or 3 years from baseline, because there were uncertainties regarding changes in the levels of inflammatory markers over approximately 7.5 years of follow-up. Fifth, we fitted a competing-risk model treating death as a competing risk. Finally, we explored an additional model adjusting for causes of CKD and years since CKD diagnosis, although a substantial number of participants had unknown causes or duration of CKD (e.g., 40% for the causes of CKD) (Table 1).
Table 1.
Baseline Characteristics According to Quartile of Interleukin-6 Concentration, Chronic Renal Insufficiency Cohort Study, United States, 2003–2013
| Characteristic | Overall (n = 3,597) | IL-6 Concentration, pg/mL | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| <1.15 (n = 899) | 1.15–1.86 (n = 897) | 1.87–3.08 (n = 901) | ≥3.09 (n = 900) | |||||||
| No. | % | No. | % | No. | % | No. | % | No. | % | |
| Agea, years | 57.9 (10.9) | 54.6 (11.6) | 57.9 (10.7) | 59.8 (10.2) | 59.2 (10.3) | |||||
| Male sex | 1,979 | 55.0 | 517 | 57.5 | 481 | 53.6 | 496 | 55.0 | 485 | 53.9 |
| Race/ethnicity | ||||||||||
| Non-Hispanic white | 1,543 | 42.9 | 483 | 53.7 | 407 | 45.4 | 336 | 37.3 | 317 | 35.2 |
| Non-Hispanic black | 1,476 | 41.0 | 295 | 32.8 | 348 | 38.8 | 401 | 44.5 | 432 | 48.0 |
| Other | 578 | 16.1 | 121 | 13.5 | 142 | 15.8 | 164 | 18.2 | 151 | 16.8 |
| Body mass indexa,b | 32.1 (7.8) | 28.9 (5.7) | 31.7 (6.9) | 33.5 (7.8) | 34.4 (9.2) | |||||
| Current smoking | 453 | 12.6 | 87 | 9.7 | 114 | 12.7 | 101 | 11.2 | 151 | 16.8 |
| Alcohol use | 2,277 | 63.3 | 674 | 75.0 | 585 | 65.2 | 523 | 58.0 | 495 | 55.0 |
| Comorbidity | ||||||||||
| Hypertension | 3,094 | 86.0 | 675 | 75.1 | 778 | 86.7 | 817 | 90.7 | 824 | 91.6 |
| Diabetes | 1,727 | 48.0 | 282 | 31.4 | 421 | 46.9 | 510 | 56.6 | 514 | 57.1 |
| Cardiovascular disease | 1,185 | 32.9 | 168 | 18.7 | 268 | 29.9 | 352 | 39.1 | 397 | 44.1 |
| Cancer | 253 | 7.0 | 52 | 5.8 | 66 | 7.4 | 75 | 8.3 | 60 | 6.7 |
| COPD | 109 | 3.0 | 18 | 2.0 | 21 | 2.3 | 27 | 3.0 | 43 | 4.8 |
| Steroid use | 359 | 10.0 | 86 | 9.6 | 82 | 9.1 | 78 | 8.7 | 113 | 12.6 |
| Cause of CKD | ||||||||||
| Diabetes | 928 | 25.8 | 155 | 17.2 | 218 | 24.3 | 274 | 30.4 | 281 | 31.2 |
| Hypertension | 587 | 16.3 | 146 | 16.2 | 156 | 17.4 | 138 | 15.3 | 147 | 16.3 |
| Other | 507 | 14.1 | 159 | 17.7 | 142 | 15.8 | 111 | 12.3 | 95 | 10.6 |
| Unknown | 1,575 | 43.8 | 439 | 48.8 | 381 | 42.5 | 378 | 42.0 | 377 | 41.9 |
| Years since CKD diagnosis | ||||||||||
| <1 year | 1,255 | 34.9 | 321 | 35.7 | 316 | 35.2 | 315 | 35.0 | 303 | 33.7 |
| 1–4 years | 1,233 | 34.3 | 297 | 33.0 | 286 | 31.9 | 311 | 34.5 | 339 | 37.7 |
| ≥5 years | 776 | 21.6 | 200 | 22.2 | 214 | 23.9 | 176 | 19.5 | 186 | 20.7 |
| Unknown | 333 | 9.3 | 81 | 9.0 | 81 | 9.0 | 99 | 11.0 | 72 | 8.0 |
| eGFR, mL/min/1.73 m2 | ||||||||||
| ≥60 | 537 | 14.9 | 256 | 28.5 | 125 | 13.9 | 84 | 9.3 | 72 | 8.0 |
| 45–59 | 1,101 | 30.6 | 312 | 34.7 | 305 | 34.0 | 264 | 29.3 | 220 | 24.4 |
| 30–44 | 1,303 | 36.2 | 243 | 27.0 | 314 | 35.0 | 371 | 41.2 | 375 | 41.7 |
| <30 | 656 | 18.2 | 88 | 9.8 | 153 | 17.1 | 182 | 20.2 | 233 | 25.9 |
| uACR, mg/g | ||||||||||
| <10 | 986 | 27.4 | 367 | 40.8 | 264 | 29.4 | 203 | 22.5 | 152 | 16.9 |
| 10–29 | 571 | 15.9 | 150 | 16.7 | 134 | 14.9 | 141 | 15.6 | 146 | 16.2 |
| 30–299 | 950 | 26.4 | 204 | 22.7 | 229 | 25.5 | 256 | 28.4 | 261 | 29.0 |
| ≥300 | 1,090 | 30.3 | 178 | 19.8 | 270 | 30.1 | 301 | 33.4 | 341 | 37.9 |
Abbreviations: CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate; uACR, urinary albumin-to-creatinine ratio.
a Values are expressed as mean (standard deviation).
b Weight (kg)/height (m)2.
Subgroup analyses were performed for prespecified groups of age (≥65 vs. <65 years), sex (men vs. women), race (non-Hispanic black vs. other), diabetes (yes vs. no), hypertension (yes vs. no), history of cardiovascular disease (yes vs. no), eGFR (in mL/min/1.73 m2: ≥60, 45–59, 30–44, and <30), and uACR (in mg/g: <10, 10–29, 30–299, and ≥300). Interaction was assessed using the log-likelihood test. A 2-sided P value of <0.05 was used to test statistical significance. All statistical analyses were performed using Stata, version 15 (StataCorp LLC, College Station, Texas).
RESULTS
Baseline characteristics
In the overall sample, mean age was 57.9 years, 55.0% were male, 42.9% were non-Hispanic white, and 41.0% were non-Hispanic black. Table 1 shows baseline characteristics according to quartile of IL-6. Those with higher levels of IL-6 were more likely to be older, of non-Hispanic black race, and current smokers; have higher body mass index and greater prevalence of hypertension, diabetes, cardiovascular disease, and chronic obstructive pulmonary disease; and have lower eGFR and higher uACR (Table 1). Those with higher levels of TNF-α had mostly similar characteristics to those with higher levels of IL-6 (Web Table 2). The greater prevalence of lower eGFR and higher uACR was also observed among those with higher levels of IL-1RA but was less evident for TGF-β (Web Tables 3 and 4).
Table 2.
Incidence Rates and Adjusted Hazard Ratios for Risk of Hospitalization With Major Infection, Chronic Renal Insufficiency Cohort Study, United States, 2003–2013
| Exposure | No. | No. of Events | IR Per 1,000 Person-Years | 95% CI | Model 1 a | Model 2 b | Model 3 c | |||
|---|---|---|---|---|---|---|---|---|---|---|
| Adjusted HR | 95% CI | Adjusted HR | 95% CI | Adjusted HR | 95% CI | |||||
| IL-6, pg/mL | ||||||||||
| Categorical | ||||||||||
| <1.15 | 899 | 207 | 30.4 | 26.5, 34.9 | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent |
| 1.15–1.86 | 897 | 300 | 49.6 | 44.3, 55.5 | 1.49 | 1.24, 1.78 | 1.18 | 0.98, 1.42 | 1.15 | 0.96, 1.39 |
| 1.87–3.08 | 901 | 361 | 66.7 | 60.2, 74.0 | 1.97 | 1.65, 2.35 | 1.40 | 1.17, 1.69 | 1.35 | 1.12, 1.62 |
| ≥3.09 | 900 | 422 | 90.0 | 81.8, 99.0 | 2.54 | 2.13, 3.01 | 1.63 | 1.36, 1.96 | 1.53 | 1.27, 1.84 |
| Continuous | ||||||||||
| Per log (SD) increase | 1.32 | 1.26, 1.38 | 1.19 | 1.13, 1.26 | 1.14 | 1.04, 1.24 | ||||
| TNF-α, pg/mL | ||||||||||
| Categorical | ||||||||||
| <1.50 | 845 | 193 | 30.5 | 26.5, 35.2 | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent |
| 1.50–2.19 | 887 | 279 | 46.4 | 41.2, 52.1 | 1.46 | 1.22, 1.76 | 1.18 | 0.98, 1.43 | 1.17 | 0.97, 1.41 |
| 2.20–3.19 | 914 | 390 | 72.6 | 65.8, 80.2 | 2.29 | 1.92, 2.74 | 1.58 | 1.31, 1.89 | 1.52 | 1.26, 1.83 |
| ≥3.20 | 951 | 428 | 81.5 | 74.2, 89.7 | 2.51 | 2.10, 2.99 | 1.62 | 1.34, 1.96 | 1.52 | 1.26, 1.85 |
| Continuous | ||||||||||
| Per log (SD) increase | 1.31 | 1.24, 1.37 | 1.19 | 1.13, 1.27 | 1.12 | 1.01, 1.24 | ||||
| IL-1RA, pg/mL | ||||||||||
| Categorical | ||||||||||
| <384.3 | 899 | 281 | 46.6 | 41.4, 52.4 | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent |
| 384.3–707.5 | 899 | 320 | 56.1 | 50.3, 62.6 | 1.1 | 0.93, 1.29 | 1.01 | 0.86, 1.19 | 0.93 | 0.79, 1.10 |
| 707.6–1,527.6 | 899 | 333 | 57.5 | 51.6, 64.0 | 1.13 | 0.96, 1.33 | 0.96 | 0.81, 1.13 | 0.89 | 0.75, 1.05 |
| ≥1,527.7 | 900 | 356 | 65.6 | 59.2, 72.8 | 1.14 | 0.97, 1.35 | 1.02 | 0.86, 1.20 | 0.94 | 0.79, 1.11 |
| Continuous | ||||||||||
| Per log (SD) increase | 1.06 | 1.01, 1.13 | 1.02 | 0.96, 1.08 | 1.08 | 0.93, 1.26 | ||||
| TGF-β, ng/mL | ||||||||||
| Categorical | ||||||||||
| <6.36 | 899 | 308 | 51.9 | 46.4, 58.0 | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent |
| 6.36–10.88 | 899 | 344 | 59.9 | 53.9, 66.5 | 1.00 | 0.85, 1.17 | 1.02 | 0.87, 1.20 | 1.00 | 0.85, 1.17 |
| 10.89–17.82 | 898 | 320 | 57.1 | 51.2, 63.7 | 0.94 | 0.79, 1.11 | 1.03 | 0.88, 1.22 | 1.01 | 0.86, 1.20 |
| ≥17.83 | 901 | 318 | 56.1 | 50.3, 62.7 | 0.90 | 0.76, 1.06 | 0.96 | 0.81, 1.14 | 0.93 | 0.79, 1.11 |
| Continuous | ||||||||||
| Per log (SD) increase | 0.95 | 0.89, 1.01 | 0.99 | 0.93, 1.05 | 1.00 | 0.88, 1.13 | ||||
Abbreviations: CI, confidence interval; HR, hazard ratio; IL-1RA, interleukin-1 receptor antagonist; IL-6, interleukin-6; IR, incidence rate; SD, standard deviation; TGF-β, transforming growth factor-β; TNF-α, tumor necrosis factor-α.
a Stratified by center and adjusted for age, sex, and race.
b Model 1 plus adjustment for body mass index, smoking status, alcohol consumption, hypertension, diabetes, past medical history (chronic obstructive pulmonary disease, cancer, and cardiovascular disease), steroids, estimated glomerular rate, and urinary albumin-to-creatinine ratio.
c Model 2 plus adjustment for the inflammatory markers.
Inflammatory markers and incident hospitalization with major infection
During a median follow-up of 7.5 years, 35.9% (n = 1,290) had incident hospitalization with major infection (crude incidence rate = 56.2 per 1,000 person-years). Figure 1 shows the cumulative incidence of hospitalization with major infection according to quartiles of IL-6, TNF-α, IL-1RA, and TGF-β using the Kaplan-Meier method. The risks were higher with higher levels of IL-6 and TNF-α, with an approximately 3-fold risk gradient across quartiles. A similar pattern was observed for IL-1RA, but the risk gradient was less evident. The risk was mostly similar across the quartiles of TGF-β. Similar patterns were observed when accounting for death as a competing risk event (Web Figure 2).
Figure 1.
Kaplan-Meier curves for the cumulative incidence of hospitalization with major infection according to quartiles of interleukin-6 (IL-6) (A); tumor necrosis factor (TNF-α) (B); interleukin 1 receptor antagonist (IL-1RA) (C); and tumor growth factor-β (TGF-β) (D), Chronic Renal Insufficiency Cohort Study, United States, 2003–2013.
In the age-, sex-, and race-adjusted Cox proportional hazard models, individuals with IL-6 or TNF-α levels in the highest quartile had an approximately 2.5-fold higher risk of hospitalization with major infection compared with those in the lowest quartile (hazard ratio = 2.54 (95% confidence interval (CI): 2.13, 3.01) for IL-6 and 2.51 (95% CI: 2.10, 2.99) for TNF-α) (model 1 in Table 2). We did not observe any significant associations for IL-1RA or TGF-β. After adjusting for other confounders, including eGFR and uACR, the association remained significant and consistent for IL-6 and TNF-α (model 2). Additional adjustment for remaining inflammatory markers (e.g., adjusted for TNF-α, IL-1RA, and TGF-β for analysis of IL-6) did not materially alter the results for IL-6 or TNF-α (model 3).
A similar pattern for the association of IL-6 and TNF-α with risk of infection was confirmed when modeled as the restricted cubic spline with covariates in model 2 (Figure 2). With the 5th percentile as the referent, the hazard ratios at the 95th percentile were 2.11 (95% CI: 1.68, 2.66) for IL-6 and 1.88 (95% CI: 1.51, 2.33) for TNF-α. The associations of IL-1RA and TGF-β were not significant at any level, although the risk tended to be slightly higher at high levels of IL-1RA. Similar results were observed when the inflammatory markers were modeled as penalized splines (Web Figure 3).
Figure 2.
Histogram and hazard ratios for hospitalization with infection according to plasma levels of interleukin-6 (IL-6) (A); tumor necrosis factor (TNF-α) (B); interleukin 1 receptor antagonist (IL-1RA) (C); and tumor growth factor-β (TGF-β) (D), modeled as cubic splines, Chronic Renal Insufficiency Cohort Study, United States, 2003–2013. Vertical lines indicate 5th (referent) and 95th percentiles. Solid lines represent point estimates. Dotted lines represent the corresponding 95% confidence intervals. Models were stratified by center and adjusted for age, sex, race, body mass index, smoking status, alcohol consumption, hypertension, diabetes, past medical history (chronic obstructive pulmonary disease, cancer, and cardiovascular disease), steroids, estimated glomerular filtration rate, and urinary albumin-to-creatinine ratio.
When the association of IL-6 and TNF-α was assessed separately for type-specific infection of urinary tract infections (613 cases), pneumonia (540 cases), cellulitis and osteomyelitis (527 cases), and bacteremia and sepsis (485 cases), the hazard ratios were significant and consistent, with similar hazard ratios across types of infection (ranging from 1.1 to 1.3 per 1-standard-deviation increment of log-transformed IL-6 and TNF-α levels) (Web Table 5). The associations were consistent when restricting to infections as the principal diagnosis, censoring for incident cardiovascular disease or end-stage renal disease, excluding events that occurred within 1 or within 3 years of follow-up, limiting follow-up to the first 1 or 3 years, or treating death as a competing risk (Web Table 6). The associations were consistent when adjusting for the causes and duration of CKD in addition to the covariates in model 2 (per log increase, hazard ratio = 1.20 (95% CI: 1.13, 1.26) for IL-6 and 1.20 (95% CI: 1.13, 1.27) for TNF-α). In subgroup analysis, there was no statistically significant interaction by age, sex, body mass index, diabetes, hypertension, history of cardiovascular disease, eGFR, or uACR for either IL-6 or TNF-α (P for interaction for all > 0.05) (Figure 3). The association was significantly weaker among non-Hispanic blacks compared with other racial groups for IL-6 (P for interaction < 0.001) but not for TNF-α (P for interaction = 0.28).
Figure 3.
Subgroup analysis for adjusted hazard ratios of hospitalization with infection for interleukin-6 (IL-6) (A) and tumor necrosis factor (TNF-α) (B), Chronic Renal Insufficiency Cohort Study, United States, 2003–2013. Circles represent point estimates for hazard ratios per every standard-deviation increment in log-transformed IL-6 or TNF-α. Models were stratified by center and adjusted for age, sex, race, body mass index, smoking status, alcohol consumption, hypertension, diabetes, past medical history (chronic obstructive pulmonary disease, cancer, and cardiovascular disease), steroids, estimated glomerular filtration rate (eGFR), and urinary albumin-to-creatinine ratio (uACR). Body mass index was calculated as weight (kg)/height (m)2. Horizontal lines represent the range for 95% confidence interval. CVD, cardiovascular disease.
Cross-category of IL-6 and TNF-α and incident hospitalization with major infection
To analyze the combined association of IL-6 and TNF-α, we assessed the risk of hospitalization with infection in the cross-category according to the quartiles of IL-6 and TNF-α. Table 3 shows age-, sex-, race-adjusted incidence rates of hospitalization with major infection. Overall, there was a multiplicative association of IL-6 and TNF-α with the risk of hospitalization with infection. The incidence rate was approximately 20 per 1,000 person-years when both IL-6 and TNF-α levels were in their bottom quartile, but it increased in a graded fashion with higher levels of IL-6 and TNF-α. The group with both IL-6 and TNF-α in the top quartiles had incidence rates exceeding 100 per 1,000 person-years. The results were consistent in multivariable Cox modeling accounting for other confounders (Web Table 7). Compared with the lowest quartile of IL-6 and TNF-α, the highest quartile of IL-6 and TNF-α was associated with a 2.5-fold increased risk of major infection (hazard ratio = 2.52, 95% CI: 1.82, 3.50).
Table 3.
Age-, Sex-, and Race-Adjusted Incidence Rates for Hospitalization With Infection, Chronic Renal Insufficiency Cohort Study, United States, 2003–2013
| IL-6 Concentration, pg/mL | TNF-α Concentration, pg/mL | |||||||
|---|---|---|---|---|---|---|---|---|
| <1.50 | 1.50–2.19 | 2.20–3.19 | ≥3.20 | |||||
| IR per 1,000 Person-Years | 95% CI | IR per 1,000 Person-Years | 95% CI | IR per 1,000 Person-Years | 95% CI | IR per 1,000 Person-Years | 95% CI | |
| <1.15 | 19.2 | 14.7, 25.0 | 32.3 | 25.2, 41.3 | 45.7 | 35.0, 59.7 | 43.4 | 31.1, 60.4 |
| 1.16–1.86 | 26.2 | 19.4, 35.3 | 46.1 | 36.9, 57.6 | 71.1 | 58.6, 86.1 | 59.9 | 47.8, 75.2 |
| 1.87–3.08 | 50.3 | 38.7, 65.4 | 55.9 | 44.9, 69.8 | 68.6 | 56.2, 83.6 | 88.0 | 74.0, 104.6 |
| ≥3.09 | 55.9 | 40.9, 76.6 | 61.4 | 47.6, 79.2 | 103.5 | 87.4, 122.6 | 109.5 | 95.0, 126.3 |
Abbreviations: IL-6, interleukin-6; IR, incidence rate; TNF-α, tumor necrosis factor-α.
DISCUSSION
In the present cohort study of 3,597 individuals with a wide range of CKD stages, higher levels of IL-6 and TNF-α were significantly and independently associated with increased risk of hospitalization with major infections. Significant association was observed across different types of infection, including pneumonia, urinary tract infections, cellulitis and osteomyelitis, and bacteremia and sepsis. The associations were also robust to sensitivity analyses and consistent across subgroups of age, sex, race, body mass index, diabetes status, hypertension status, history of cardiovascular disease, eGFR, and uACR. The association of IL-1RA with infection risk was significant in demographically adjusting models but was no longer significant after additionally adjusting for other confounders. We did not see any significant associations between TGF-β levels and infection.
Our findings are consistent with previous studies showing a positive association of IL-6 and TNF-α with risk of pneumonia or sepsis in selected high-risk populations defined by older age or human immunodeficiency virus infection (15–17). We extended these findings to a CKD population and included several key demographic and clinical subgroup analyses (e.g., sex, race, body mass index, diabetes status). We also demonstrated that these inflammatory markers were consistently associated with a broad range of infections, including urinary tract infections as well as cellulitis and osteomyelitis. Furthermore, we confirmed the consistent associations both when infection diagnosis was principal and when it was secondary, suggesting that our findings were not explained merely by the high risks of hospitalization associated with high levels of IL-6 and TNF-α, which could result in infection complications. Importantly, we evaluated 4 inflammatory markers and observed evident dose-dependent associations of IL-6 and TNF-α (but not IL-1RA and TGF-β) with infection, suggesting their unique involvement in the susceptibility to infection.
Several mechanisms can help explain the link between high levels of IL-6 and TNF-α and increased risk of infection. IL-6 and TNF-α are pleiotropic proinflammatory cytokines that play an essential role in both innate and acquired immune responses (8, 9). However, animal studies have suggested that enhanced expression of these inflammatory markers could impair the function of neutrophil recruitment, macrophage activation, and macrophage phagocytosis, and could augment growth of the pathogens (10–12). IL-6 and TNF-α are also known to induce the expression of reactive oxygen species. Some evidence shows that imbalanced reactive oxygen species could result in the reduced clearance of bacteria and viruses (23). In addition, the interaction of IL-6 and TNF-α with endothelial cells might play a role in the migration, adhesion, and vascular permeability of circulating lymphocytes and leukocytes. Furthermore, CKD is strongly related to older age, and proinflammatory changes are suggested to play some roles in the age-related impaired innate and adaptive immune function (e.g., immunosenescence) (24).
We observed a weaker association of IL-6 with the risk of infection in black participants compared with participants of other race/ethnicity. Although we are not sure of potential mechanisms behind this observation, among black participants, a shallow risk gradient was especially evident at higher levels of IL-6 (e.g., >2.5 pg/mL) (Web Figure 4). Nonetheless, because we assessed several subgroups without an a priori hypothesis, our subgroup analysis should be considered hypothesis-generating. Importantly, the interaction was quantitative but not qualitative, and higher levels of IL-6 were positively associated with the risk of infection in both black and nonblack participants.
In contrast to our hypothesis, we did not observe significant associations for IL-1RA or TGF-β, which deserves some discussion. Cytokines are mostly bound to their receptors immediately after their release (25), and therefore, circulating cytokines in the blood might not necessarily reflect their actual activities in organs. Indeed, elevated concentrations of localized IL-1RA and TGF-β have been previously reported in the joints of patients with rheumatoid arthritis (26) and the colon of patients with inflammatory bowel disease (27). Also, it is possible that signaling of IL-1RA or TGF-β is regulated mainly by the modulation of their receptor or downstream pathways and thus not reflected by their blood concentrations (28, 29). Finally, antiinflammatory markers are often released in response to the proinflammatory cytokines (14), and therefore relative abundance might be of more importance. Nonetheless, future studies are needed to explore the underlying mechanisms of the different patterns of inflammatory markers and their associations with infection risk.
Our study has limitations. First, we do not know the reason for the high levels of inflammatory markers and cannot distinguish between chronic and acute or recent inflammatory processes. However, the results were consistent when excluding events that occurred within 1 or within 3 years of follow-up. Second, our outcome ascertainment relied on hospital-assigned ICD-9-CM codes, which might be subject to misclassification. However, previous studies reported overall high positive predictive value (>90%) for diagnosis of acute infection using ICD-9 codes (30, 31). Third, although we did our best to account for confounding, several other potential confounders were not available in CRIC (e.g., immune-related comorbidities such as rheumatoid arthritis and previous history of hospitalized and nonhospitalized infections). Fourth, types of pathogens that caused the infections were not available in CRIC. Fifth, our approach of complete-case analyses (i.e., excluding approximately 9% of participants, who had missing values) might have led to a selected study population. However, we confirmed comparable baseline characteristics between those who were included in and excluded from the present study. More importantly, this selection happened prior to outcome occurrence, and thus overall it is unlikely that our selected population has resulted in selection bias (32). Sixth, because inflammatory markers were assessed once at baseline, we were unable to evaluate them as time-varying exposures. Our study also has several strengths. First, the population is large and diverse. Second, the follow-up period is long, with a median of 7.5 years. Third, there was considerable variation in levels of the inflammatory markers.
The present study has important research and clinical implications. Our results offer proof of concept for future studies investigating inflammatory pathways that involve IL-6 and TNF-α increasing susceptibility to infection among individuals with CKD. Of note, both IL-6 and TNF-α are essential components of the immune system, and the inhibition of a single pathway would likely increase the risk of infection as shown in the clinical trials of monoclonal antibodies to TNF-α or IL-6 (33, 34). Nonetheless, understanding the contribution of inflammation pathways to the vulnerability to infection in CKD could help us identify a novel therapeutic target for reducing the risk of infection, as has been explored in the prevention of cardiovascular disease (35). Also, some inflammatory markers, such as IL-6 and TNF-α, might be useful for identifying CKD patients who are at a particularly high risk of infection. Such risk stratification could help the targeted application of procedures to prevent infection. In this context, although the inflammatory markers tested in our study are currently not routinely measured in clinical practice, previous studies reported the usefulness of IL-6 and TNF-α in risk stratification for some outcomes, such as mortality (22, 36, 37). Importantly, despite clinical guidelines, patients with CKD are not appropriately receiving recommended vaccinations (38). Thus, such a risk-centered approach might help targeted promotion for receiving evidence-based vaccinations.
In conclusion, among individuals with CKD, higher levels of IL-6 and TNF-α, but not necessarily IL-1RA or TGF-β, were associated with increased risk of hospitalization with major infections independently of potential confounders including eGFR and uACR. These findings suggest the relevance of IL-6 and TNF-α in the increased susceptibility to infection in CKD. Future studies are needed to explore the underlying mechanisms of the different patterns across inflammatory markers in their associations with infection risk.
Supplementary Material
ACKNOWLEDGMENTS
Author affiliations: Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland (Junichi Ishigami, Jeanne Charleston, Lawrence J. Appel, Kunihiro Matsushita); Department of Nephrology and Hypertension, Cleveland Clinic Foundation, Cleveland, Ohio (Jonathan Taliercio); Department of Biostatistics, Epidemiology, and Informatics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania (Harold I. Feldman, Raymond Townsend, Debbie L. Cohen); Division of Nephrology and Hypertension, Feinberg School of Medicine, Northwestern University, Chicago, Illinois (Anand Srivastava, Tamara Isakova); Division of Nephrology, MetroHealth Medical Center, Cleveland, Ohio (Edward Horwitz); Division of Nephrology, University of Michigan, Ann Arbor, Michigan (Panduranga Rao); Department of Epidemiology and Public Health, School of Medicine, University of Maryland, Baltimore, Maryland (Jeffrey C. Fink); Division of Nephrology, University of Illinois, Chicago, Illinois (Ana C. Ricardo); Division of Nephrology and Hypertension, School of Medicine, Wayne State University, Detroit, Michigan (James Sondheimer); Division of Nephrology, School of Medicine, Johns Hopkins University, Baltimore, Maryland; (Teresa K. Chen); and Division of Nephrology, School of Medicine, Duke University, Durham, North Carolina (Myles Wolf).
J.I. was supported by the National Heart, Lung, and Blood Institute (grant T32HL007024). This research was also supported by the National Institute of Diabetes and Digestive and Kidney Diseases (grant R01DK1100870 to T.I.) Funding for the Chronic Renal Insufficiency Cohort Study was obtained under a cooperative agreement from the National Institute of Diabetes and Digestive and Kidney Diseases (grants U01DK060990, U01DK060984, U01DK061022, U01DK061021, U01DK061028, U01DK060980, U01DK060963, and U01DK060902). In addition, this work was supported in part by the Perelman School of Medicine at the University of Pennsylvania (Clinical and Translational Science Award NIH/NCATS UL1TR000003), Johns Hopkins University (grant UL1 TR-000424), University of Maryland (grant GCRC M01 RR-16500), Clinical and Translational Science Collaborative of Cleveland, National Center for Advancing Translational Sciences (grant UL1TR000439) component of the National Institutes of Health and its Roadmap for Medical Research, Michigan Institute for Clinical and Health Research (grant UL1TR000433), University of Illinois at Chicago (CTSA UL1RR029879), Tulane Center of Biomedical Research Excellence for Clinical and Translational Research in Cardiometabolic Diseases (grant P20 GM109036), and Kaiser Permanente (NIH/NCRR UCSF-CTSI UL1 RR-024131).
This study was presented in abstract form at the 2018 American Society of Nephrology Kidney Week, October 23–28, 2018, San Diego, California.
The Chronic Renal Insufficiency Cohort Study Investigators include Alan S. Go, MD (Kaiser Permanente Division of Research, Oakland, California); Jiang He, MD, PhD (Tulane University); James P. Lash, MD (University of Illinois); and Mahboob Rahman, MD (Case Western Reserve University School of Medicine).
The sponsors played no role in the design and conduct of the study; the collection, management, analysis, and interpretation of the data; the preparation, review, and approval of the manuscript; or the decision to submit the manuscript for publication. The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Conflicts of interest: none declared.
REFERENCES
- 1. Saran R, Li Y, Robinson B, et al. US Renal Data System 2014 annual data report: epidemiology of kidney disease in the United States. Am J Kidney Dis. 2015;65(1 suppl 1):S1–S305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Xu H, Gasparini A, Ishigami J, et al. eGFR and the risk of community-acquired infections. Clin J Am Soc Nephrol. 2017;12(9):1399–1408. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Ishigami J, Grams ME, Chang AR, et al. CKD and risk for hospitalization with infection: the Atherosclerosis Risk in Communities (ARIC) study. Am J Kidney Dis. 2017;69(6):752–761. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Powe NR, Jaar B, Furth SL, et al. Septicemia in dialysis patients: incidence, risk factors, and prognosis. Kidney Int. 1999;55(3):1081–1090. [DOI] [PubMed] [Google Scholar]
- 5. Ishigami J, Matsushita K. Clinical epidemiology of infectious disease among patients with chronic kidney disease. Clin Exp Nephrol. 2019;23(4):437–447. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Gupta J, Mitra N, Kanetsky PA, et al. Association between albuminuria, kidney function, and inflammatory biomarker profile in CKD in CRIC. Clin J Am Soc Nephrol. 2012;7(12):1938–1946. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Shlipak MG, Fried LF, Crump C, et al. Elevations of inflammatory and procoagulant biomarkers in elderly persons with renal insufficiency. Circulation. 2003;107(1):87–92. [DOI] [PubMed] [Google Scholar]
- 8. Rose-John S, Winthrop K, Calabrese L. The role of IL-6 in host defence against infections: immunobiology and clinical implications. Nat Rev Rheumatol. 2017;13(7):399–409. [DOI] [PubMed] [Google Scholar]
- 9. Kalliolias GD, Ivashkiv LB. TNF biology, pathogenic mechanisms and emerging therapeutic strategies. Nat Rev Rheumatol. 2016;12(1):49–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Mason CM, Dobard E, Summer WR, et al. Intraportal lipopolysaccharide suppresses pulmonary antibacterial defense mechanisms. J Infect Dis. 1997;176(5):1293–1302. [DOI] [PubMed] [Google Scholar]
- 11. Lee JH, Del Sorbo L, Khine AA, et al. Modulation of bacterial growth by tumor necrosis factor-alpha in vitro and in vivo. Am J Respir Crit Care Med. 2003;168(12):1462–1470. [DOI] [PubMed] [Google Scholar]
- 12. Bermudez LE, Wu M, Petrofsky M, et al. Interleukin-6 antagonizes tumor necrosis factor-mediated mycobacteriostatic and mycobactericidal activities in macrophages. Infect Immun. 1992;60(10):4245–4252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Arend WP, Malyak M, Guthridge CJ, et al. Interleukin-1 receptor antagonist: role in biology. Annu Rev Immunol. 1998;16:27–55. [DOI] [PubMed] [Google Scholar]
- 14. Sanjabi S, Zenewicz LA, Kamanaka M, et al. Anti-inflammatory and pro-inflammatory roles of TGF-beta, IL-10, and IL-22 in immunity and autoimmunity. Curr Opin Pharmacol. 2009;9(4):447–453. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Wang HE, Shapiro NI, Griffin R, et al. Inflammatory and endothelial activation biomarkers and risk of sepsis: a nested case-control study. J Crit Care. 2013;28(5):549–555. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Bjerk SM, Baker JV, Emery S, et al. Biomarkers and bacterial pneumonia risk in patients with treated HIV infection: a case-control study. PLoS One. 2013;8(2):e56249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Yende S, Tuomanen EI, Wunderink R, et al. Preinfection systemic inflammatory markers and risk of hospitalization due to pneumonia. Am J Respir Crit Care Med. 2005;172(11):1440–1446. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Feldman HI, Appel LJ, Chertow GM, et al. The Chronic Renal Insufficiency Cohort (CRIC) study: design and methods. J Am Soc Nephrol. 2003;14(7 suppl 2):S148–S153. [DOI] [PubMed] [Google Scholar]
- 19. Christensen KL, Holman RC, Steiner CA, et al. Infectious disease hospitalizations in the United States. Clin Infect Dis. 2009;49(7):1025–1035. [DOI] [PubMed] [Google Scholar]
- 20. Lash JP, Go AS, Appel LJ, et al. Chronic Renal Insufficiency Cohort (CRIC) study: baseline characteristics and associations with kidney function. Clin J Am Soc Nephrol. 2009;4(8):1302–1311. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Inker LA, Schmid CH, Tighiouart H, et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med. 2012;367(1):20–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Volpato S, Guralnik JM, Ferrucci L, et al. Cardiovascular disease, interleukin-6, and risk of mortality in older women: the Women's Health and Aging Study. Circulation. 2001;103(7):947–953. [DOI] [PubMed] [Google Scholar]
- 23. Paiva CN, Bozza MT. Are reactive oxygen species always detrimental to pathogens? Antioxid Redox Signal. 2014;20(6):1000–1037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Fulop T, Larbi A, Dupuis G, et al. Immunosenescence and inflamm-aging as two sides of the same coin: friends or foes? Front Immunol. 2017;8:1960. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Arend WP. The balance between IL-1 and IL-1Ra in disease. Cytokine Growth Factor Rev. 2002;13(4–5):323–340. [DOI] [PubMed] [Google Scholar]
- 26. Barrera P, van der Laken CJ, Boerman OC, et al. Radiolabelled interleukin-1 receptor antagonist for detection of synovitis in patients with rheumatoid arthritis. Rheumatology. 2000;39(8):870–874. [DOI] [PubMed] [Google Scholar]
- 27. Del Zotto B, Mumolo G, Pronio AM, et al. TGF-beta1 production in inflammatory bowel disease: differing production patterns in Crohn's disease and ulcerative colitis. Clin Exp Immunol. 2003;134(1):120–126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Garlanda C, Dinarello CA, Mantovani A. The interleukin-1 family: back to the future. Immunity. 2013;39(6):1003–1018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Huang F, Chen YG. Regulation of TGF-beta receptor activity. Cell Biosci. 2012;2:9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Schneeweiss S, Robicsek A, Scranton R, et al. Veteran’s affairs hospital discharge databases coded serious bacterial infections accurately. J Clin Epidemiol. 2007;60(4):397–409. [DOI] [PubMed] [Google Scholar]
- 31. Drahos J, Vanwormer JJ, Greenlee RT, et al. Accuracy of ICD-9-CM codes in identifying infections of pneumonia and herpes simplex virus in administrative data. Ann Epidemiol. 2013;23(5):291–293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. White IR, Carlin JB. Bias and efficiency of multiple imputation compared with complete-case analysis for missing covariate values. Stat Med. 2010;29(28):2920–2931. [DOI] [PubMed] [Google Scholar]
- 33. Smolen JS, Beaulieu A, Rubbert-Roth A, et al. Effect of interleukin-6 receptor inhibition with tocilizumab in patients with rheumatoid arthritis (OPTION study): a double-blind, placebo-controlled, randomised trial. Lancet. 2008;371(9617):987–997. [DOI] [PubMed] [Google Scholar]
- 34. Present DH, Rutgeerts P, Targan S, et al. Infliximab for the treatment of fistulas in patients with Crohn's disease. N Engl J Med. 1999;340(18):1398–1405. [DOI] [PubMed] [Google Scholar]
- 35. Ridker PM, Everett BM, Thuren T, et al. Antiinflammatory therapy with canakinumab for atherosclerotic disease. N Engl J Med. 2017;377(12):1119–1131. [DOI] [PubMed] [Google Scholar]
- 36. Munoz Mendoza J, Isakova T, Cai X, et al. Inflammation and elevated levels of fibroblast growth factor 23 are independent risk factors for death in chronic kidney disease. Kidney Int. 2017;91(3):711–719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Amdur RL, Feldman HI, Gupta J, et al. Inflammation and progression of CKD: the CRIC Study. Clin J Am Soc Nephrol. 2016;11(9):1546–1556. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Ishigami J, Padula WV, Grams ME, et al. Cost-effectiveness of pneumococcal vaccination among patients with CKD in the United States. Am J Kidney Dis. 2019;74(1):23–35. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.