Abstract
Background
Serum C-reactive protein (CRP) is a marker of acute inflammatory response and has been associated with health outcomes in some studies. Inflammation and immune response may have potential prognostic implications for breast cancer survivors.
Methods
The Women’s Healthy Eating and Living (WHEL) Study includes 2919 early stage breast cancer survivors with serum collected 2 years post-diagnosis and follow-up for clinical outcomes over approximately 7 years. CRP concentrations were measured using high-sensitivity electrochemiluminescence assay. Outcomes, including all-cause mortality, breast cancer-specific mortality, and additional breast cancer events were oncologist verified from medical records and death certificates. Cox proportional hazards models were conducted with adjustment for potential confounding factors to generate hazard ratios (HR) and 95% confidence intervals (CI).
Results
CRP concentrations in women diagnosed with breast cancer were associated with death due to any cause, death due to breast cancer, and additional breast cancer events, after adjustment for sociodemographic and cancer characteristics (lnCRP: P<0.05 for all three outcomes). The HR for women with (versus without) acute inflammation suggests a threshold effect on overall survival, rather than a dose-response relationship (≥10.0 mg/L v <1 mg/L: HR 1.96; 95% CI, 1.22–3.13). Associations were similar for breast cancer-specific mortality (HR 1.91; 95% CI, 1.13–3.23) and any additional breast cancer-related event (HR 1.69; 95% CI, 1.17–2.43).
Conclusions
Acute inflammation status (CRP ≥10 mg/L) may be an important independent biomarker for long-term survival in breast cancer survivors.
Impact
Interventions to decrease circulating CRP concentrations in breast cancer survivors with acute inflammation may improve prognosis.
Keywords: C-reactive protein, all-cause mortality, breast cancer-specific mortality, recurrence
INTRODUCTION
An estimated 2.9 million women in the United States live with a history of breast cancer and these women are susceptible to various comorbidities, including disease recurrence or new primary cancer (1, 2). Evidence from both laboratory and epidemiologic data suggest chronic inflammation facilitates tumor growth and metastasis through the modification of tumor cell biology, activation of stromal cells in the tumor microenvironment, cancer cell invasion by the conditioning of vasculature to enhance the extravasation, engraftment and growth of micrometastases, or reactivate dormant tumors at distant sites (3–6). Cancer survivors with persistent inflammation may have an elevated risk of recurrence or new primary as a result of the effects of inflammatory processes or the presence of cancer cells that induce inflammation (7). Thus understanding the association between postdiagnosis inflammation and prognosis is a high research priority.
A number of studies have examined inflammatory markers as predictors of disease recurrence and death in cancer populations (7–12). Although several biomarkers of inflammation have been examined, including white blood cells, fibrinogen, interleukin-6, tumor necrosis factor-α, serum amyloid A, and C-reactive protein (CRP); the most consistent association with prognosis has been with CRP (2, 12–15). CRP is a nonspecific, acute-phase protein produced by the liver in response to inflammation, infection, and tissue damage (6). As one of the most sensitive acute-phase reactants, circulating concentrations of CRP increases rapidly in response to numerous pathological and disease conditions (16). While the acute-phase response is not diagnostic for any particular disease, elevated concentrations of CRP, conventionally defined as circulating concentrations of CRP ≥3 mg/L, has proven useful in the clinical setting to monitor infections and postoperative complications, to assess effectiveness of treatments on the course of a disease, and to estimate risk of future cardiovascular events (17). With improved measurement techniques, specifically high-sensitivity assays, researchers can detect minute concentrations of circulating CRP (min detectable limit 0.02 mg/mL). Advancements in measurement of CRP combined with knowledge of the role that inflammation plays in cancer development and progression has led to a growing interest in determining the clinical utility of CRP for predicting recurrence risk, for monitoring risk reduction, and for guiding preventive approaches in women previously diagnosed with breast cancer.
Longitudinal studies in women diagnosed with breast cancer have reported conflicting results in relation to inflammation and prognosis, with some studies showing an association between elevated CRP and poor prognosis (18–22) and others showing no relationship (23–25). In 2011, a meta-analysis was conducted using 10 studies (n=4502) in women diagnosed with breast cancer. Han et al. reported a pooled HR for overall survival at 1.62 (95%CI 1.20–2.18) and a higher pooled HR for cancer-specific survival (HR=2.08; 95%CI, 1.48–2.94) (26) for elevated CRP measured across the continuum of the breast cancer experience from pre- to post-diagnosis (26). It is plausible that the variation in the CRP-survival relationship seen in previous reports could be partially attributed to different cut points. Thus, it is unclear whether clinically relatable cut points for CRP values are associated with breast cancer prognosis among women diagnosed with breast cancer, specifically categories used for cardiovascular risk prediction: low risk (<1.0 mg/L), moderate risk (1.0 to 3.0 mg/l), high risk (>3 to 10 mg/L) and values indicating acute infection (≥10 mg/L) (17).
Here, we report on the prevalence of inflammatory status, using post-treatment serum concentrations of CRP measured on average 24 months post-diagnosis, and the association with all-cause and breast cancer-specific mortality and additional breast cancer events in a cohort of 2919 women following a diagnosis of invasive breast cancer (stage I to IIIA, AJCC IV classification). We also examine the influence of individual and breast cancer clinical characteristics on inflammatory status.
MATERIALS AND METHODS
Design Overview, Participants and Methods
The Women’s Healthy Eating and Living (WHEL) study was a randomized controlled trial aimed at examining the effects of a high-vegetable, low-fat diet in reducing additional breast cancer events and early death in women diagnosed with breast cancer. Study details and the intervention have been described (27). In brief, between 1995 and 2000, 3088 women were enrolled and followed through 2006. Participants were enrolled an average of 2 years post-diagnosis, were diagnosed with stage I - III invasive breast cancer, and had completed active treatment. Participants were 27 to 74 years of age and had no evidence of disease within 12 months of study enrollment. The study was performed with the approval of the institutional review boards of the University of California, San Diego and 6 other participating centers. All participants provided informed written consent.
At baseline, the mean age of participants was 53 years; 85.5% were non-Hispanic white; 84.9% had had stage I or II breast cancer; 56.5% had well or moderately differentiated tumors; 77.8% had estrogen receptor-positive and/or progesterone receptor-positive tumors; 61.6% had received radiation therapy; 69.3% had received adjuvant chemotherapy; and 61.5% reported taking anti-estrogen medication at study entry. Women included in this study were followed semiannually and had a median follow-up of 7.3 years from the time of study enrollment.
There was no intervention effect on additional breast cancer events or mortality during the 7.3-year follow-up period. (28). Accordingly, we treated the WHEL study as a cohort study.
Serum C-reactive Protein Assay
Using stored fasting serum specimens drawn a mean 23.6 months post-diagnosis (median: 21.7 months; range: 2 to 48 months), we measured serum concentrations of CRP using a high-sensitivity electrochemiluminescence assay (MesoScale Discovery, Gaithersburg, MD) at the Laboratory for Clinical Biochemistry Research, University of Vermont. The lower detection limit for this assay platform was 0.02 mg/L. For each assay, we included blinded duplicates for quality control assessment. The inter-assay coefficients of variation were between 7 to 12%.
Outcomes & Follow-Up
Women were followed for vital status from study entry until end of study, June 1, 2006. Information on outcomes, including all-cause and breast cancer-specific mortality and additional breast cancer events (defined as recurrence [85%] or new primary breast cancer [15%]) were obtained by self-report every 6 months throughout the study and confirmed by an oncologist using medical records and death certificates. Causes of death were coded using International Classification of Diseases, 9th Revision (ICD-9) codes. Approximately four percent of study participants were lost to follow-up and these were censored at date of last contact.
Anthropometrics
At the baseline clinical visit, each participant had height, weight, waist and hip measurements conducted by trained staff using standardized measurement protocols (27). Body mass index (BMI=kg/m2) was calculated from weight and height, measured to the nearest 0.1 kg and 0.1cm, respectively, with a balance beam scale and stadiometer.
Other Variables
Standardized questionnaire information was collected at baseline on medical history, demographic characteristics and lifestyle factors. At enrollment, women were considered postmenopausal if they were amenorrheic for at least 12 months and premenopausal if they reported at least one menstrual cycle during the last 3 months. Women were classified as users of anti-estrogen medication (99.7% of which was selective estrogen receptor modulators), if they reported current use at the baseline clinic visit. History of inflammatory-related conditions (e.g., cardiovascular disease, arthritis and diabetes) was self-reported at baseline. Recreational physical activity at time of study entry was computed using a self-report validated scale designed for the Women’s Health Initiative (29). Self-report of current and former smoking status was recorded and alcohol and dietary intake data were obtained via repeated 24-h dietary recalls (described elsewhere in more detail) (30). We converted physical activity into metabolic equivalent task hours/week (31), tobacco use was converted to pack-years and the multiple dietary recalls were used to characterize participant alcohol consumption (g/day), fruit and vegetable intake (servings/day), dietary fiber intake (g/d), and percentage energy from fat.
Exclusions
Among the 3088 eligible women enrolled at baseline, serum samples were available for 3023 participants. We excluded those with outcome events within 9 months of the baseline blood draw (n=85) as well as those for whom CRP was not measured successfully (n=19), resulting in a final sample size of 2919 participants.
Statistical Analysis
Our analytic goals were to describe the distribution of serum CRP in this cohort of breast cancer survivors and characterize the association of inflammatory status with all-cause mortality, breast cancer mortality, or additional breast cancer events. We modeled inflammatory status using two methods: the first classified serum CRP concentrations into four categories of inflammatory status based on CVD prediction cut points: no inflammation (<1 mg/L), low inflammation (1 to 3 mg/L), moderate inflammation (>3 to <10 mg/L) and acute inflammation (≥10 mg/L) (17); the second treated CRP as a log-transformed continuous variable to reduce skewness of the distribution. For categorical analyses, the reference group was women with no inflammation. Trend tests, which used a single ordinal indicator variable based on inflammatory status (coded 1 for no inflammation to 4 for acute inflammation) were used to assess dosage.
We employed delayed-entry Cox proportional hazards models (with entry time at baseline blood draw) to estimate the age-adjusted and multivariable adjusted hazard ratios (HRs) and 95% confidence intervals (CI) for death due to any cause, breast cancer-specific death and disease recurrence by inflammatory status (32). Time since diagnosis (years) was used as the time metric for all regression analysis. Variables were retained in the final model (model 1) if they were associated with elevated CRP status (>3 mg/L), were associated with overall mortality in women with no inflammation (CRP <1 mg/L), and had altered the risk estimate of the model containing inflammatory status plus age by at least 10%.We further adjusted for factors that improved model fit (model 2), and /or allowed comparison to the published literature (model 3) (16, 18, 19, 33). The three models were adjusted for age at diagnosis, time since diagnosis, race-ethnicity, and tumor stage and grade (model 1); model 1 covariates and BMI (model 2); model 2 covariates plus anti-estrogen medication use and ER/PR status (model 3). Additional variables considered for adjustment included intervention arm (intervention/control), menopausal status (pre- and perimenopausal/postmenopausal), treatment used (chemotherapy, radiation, or both), physical activity (MET-hr/wk), and tobacco use (pack-yr); however, adjusting for these covariates did not materially change the risk estimates for CRP (data not shown). We tested and confirmed non-violation of the proportionality assumption based on a graphical approach (log(-log) plots) (32) and the goodness-of-fit test using Schoenfeld residuals (34). Given the high proportion of deaths due to breast cancer in this cohort, we also considered whether the association between inflammation and breast cancer-specific death (79.7% of the deaths) varied according to subgroups defined by age (>55 yrs), postmenopausal status (yes/no), BMI (≥25 kg/m2), tumor stage (≥stage II), tumor hormone receptor status (ER+), and inflammatory-related illness, including cardiovascular disease (CVD) (yes/no), arthritis (yes/no) and diabetes (yes/no). Effect modification by these factors was tested by adding a product term for inflammatory status and each of the aforementioned covariates. Statistical tests were performed using Stata (version 11.1; StataCorp LP, College Station, TX) and SAS (version 9.3 Cary NC) software. All tests were two-sided and statistical significance was set at P <0.05.
RESULTS
In this cohort of 2919 women with a history of breast cancer, the distribution of serum concentrations of CRP was markedly skewed with 91.3% of survivors having concentrations less than 10 mg/L. The geometric mean for serum CRP was 1.71 mg/L (IQR, 0.67–4.24) and median was 3.83 mg/L ± 6.7 (data not shown). Here we summarize statistically significant participant and tumor characteristics associated with this biomarker of inflammation.
Serum CRP increased with age, BMI, and postmenopausal status and the distributions of CRP concentrations varied across racial-ethnic groups (Table 1). Asian women had the lowest average CRP concentration (0.91 mg/L), non-Hispanic black women had higher concentrations (3.38 mg/L), and non-Hispanic white and Hispanic women were within those bounds (1.67 and 2.13, respectively). Less healthful behaviors were associated with increased CRP, such as tobacco use, lack of physical activity, alcohol consumption, and poor dietary habits (e.g., <5 servings/d of fruits and vegetables, <20 g/d of dietary fiber intake, and ≥30% energy intake from fat). Women with more aggressive tumor characteristics, specifically tumor stage, and estrogen and progesterone receptor negative tumors (ER−/PR− v ER+ or PR+) had higher CRP values (Table 2). Women using anti-estrogen medication had lower CRP concentrations than women who did not (1.54 and 2.03 mg/L, respectively, Table 2).
Table 1.
Serum concentrations of C-reactive Protein by participant characteristics in breast cancer survivors of the Women's Healthy Eating and Living Study
| Distribution of Inflammatory Status | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| No Inflammation |
Low Inflammation |
Moderate Inflammation |
Acute Inflammation |
||||||||
| Serum CRP mg/L |
<1 mg/L | 1–3 mg/L | >3 to <10 mg/L | >10 mg/L | |||||||
| n (%) | Geometric mean |
No. | % | No. | % | No. | % | No. | % | P | |
| Total | 2919 | 1.7 | 981 | 33.6 | 933 | 31.96 | 750 | 25.69 | 255 | 8.74 | |
| Age at study entry, years | <0.001 | ||||||||||
| 20–29 | 6 (0.2) | 0.40 | 5 | 0.5 | 1 | 0.1 | 0 | 0.0 | 0 | 0.0 | |
| 30–39 | 188 (6.4) | 1.19 | 83 | 8.5 | 57 | 6.1 | 34 | 4.5 | 14 | 5.5 | |
| 40–49 | 872 (29.9) | 1.57 | 331 | 33.7 | 244 | 26.2 | 219 | 29.2 | 78 | 30.6 | |
| 50–59 | 1159 (39.7) | 1.74 | 382 | 38.9 | 382 | 40.9 | 297 | 39.6 | 98 | 38.4 | |
| 60–69 | 610 (20.9) | 2.04 | 162 | 16.5 | 218 | 23.4 | 172 | 22.9 | 58 | 22.7 | |
| 70+ | 84 (2.9) | 2.32 | 18 | 1.8 | 31 | 3.3 | 28 | 3.7 | 7 | 2.7 | |
| Race/ethnicity | <0.001 | ||||||||||
| White, non-Hispanic | 2495 (85.5) | 1.67 | 853 | 87.0 | 797 | 85.4 | 641 | 85.5 | 204 | 80.0 | |
| Black, non-Hispanic | 110 (3.8) | 3.38 | 12 | 1.2 | 40 | 4.3 | 34 | 4.5 | 24 | 9.4 | |
| Hispanic | 154 (5.3) | 2.13 | 44 | 4.5 | 50 | 5.4 | 47 | 6.3 | 13 | 5.1 | |
| Asian | 91 (3.1) | 0.91 | 49 | 5.0 | 26 | 2.8 | 15 | 2.0 | 1 | 0.4 | |
| Other | 69 (2.4) | 1.95 | 23 | 2.3 | 20 | 2.1 | 13 | 1.7 | 13 | 5.1 | |
| BMI (kg/m2) | <0.001 | ||||||||||
| <18.5, underweight | 27 (0.9) | 0.48 | 21 | 2.1 | 3 | 0.3 | 2 | 0.3 | 1 | 0.4 | |
| 18.5–24.9, normal weight | 1222 (41.9) | 0.83 | 699 | 71.3 | 363 | 38.9 | 134 | 17.9 | 26 | 10.2 | |
| 25–29.9, overweight | 905 (31.0) | 2.02 | 211 | 21.5 | 382 | 40.9 | 250 | 33.3 | 62 | 24.3 | |
| 30 to <35, obese class I | 453 (15.5) | 3.93 | 35 | 3.6 | 134 | 14.4 | 213 | 28.4 | 71 | 27.8 | |
| >35 to <40, obese class II | 200 (6.9) | 4.77 | 13 | 1.3 | 43 | 4.6 | 103 | 13.7 | 41 | 16.1 | |
| >40, obese class III | 112 (3.8) | 9.03 | 2 | 0.2 | 8 | 0.9 | 48 | 6.4 | 54 | 21.2 | |
| Waist/Hip ratio | <0.001 | ||||||||||
| <0.80 | 1525 (52.2) | 1.11 | 720 | 73.4 | 472 | 50.6 | 263 | 35.1 | 69 | 27.1 | |
| ≥0.80 | 1384 (47.4) | 2.75 | 258 | 26.3 | 454 | 48.7 | 485 | 64.7 | 185 | 72.5 | |
| Menopausal status | <0.001 | ||||||||||
| Premenopausal | 323 (11.1) | 1.16 | 148 | 15.1 | 89 | 9.5 | 65 | 8.7 | 21 | 8.2 | |
| Perimenopausal | 272 (9.3) | 1.37 | 114 | 11.6 | 77 | 8.3 | 62 | 8.3 | 19 | 7.5 | |
| Postmenopausal | 2324 (79.6) | 1.86 | 719 | 73.3 | 767 | 82.2 | 623 | 83.1 | 215 | 84.3 | |
| Lifestyle Factors | |||||||||||
| Tobacco Use (pack-years) | 0.001 | ||||||||||
| None | 1558 (53.4) | 1.64 | 540 | 55.0 | 496 | 53.2 | 390 | 52.0 | 132 | 51.8 | |
| <20 | 967 (33.1) | 1.65 | 337 | 34.4 | 312 | 33.4 | 247 | 32.9 | 71 | 27.8 | |
| ≥20 | 349 (12.0) | 2.35 | 86 | 8.8 | 112 | 12.0 | 105 | 14.0 | 46 | 18.0 | |
| Alcohol consumption, 10g/d | <0.001 | ||||||||||
| None | 1588 (54.4) | 1.95 | 482 | 49.1 | 489 | 52.4 | 456 | 60.8 | 161 | 63.1 | |
| <1 | 784 (26.9) | 1.48 | 289 | 29.5 | 255 | 27.3 | 181 | 24.1 | 59 | 23.1 | |
| ≥1 | 540 (18.5) | 1.44 | 209 | 21.3 | 187 | 20.0 | 109 | 14.5 | 35 | 13.7 | |
| Fruit/Vegetable, Servings per day | <0.001 | ||||||||||
| <5 | 1133 (38.8) | 1.99 | 329 | 33.5 | 357 | 38.3 | 326 | 43.5 | 121 | 47.5 | |
| >=5 | 1779 (61.0) | 1.55 | 651 | 66.4 | 574 | 61.5 | 420 | 56.0 | 134 | 52.5 | |
| Dietary Fiber, g/d | <0.001 | ||||||||||
| <20 | 1485 (50.9) | 2.02 | 419 | 42.7 | 483 | 51.8 | 430 | 57.3 | 153 | 60.0 | |
| >=20 | 1427 (48.9) | 1.43 | 561 | 57.2 | 448 | 48.0 | 316 | 42.1 | 102 | 40.0 | |
| % Energy from Fat | <0.001 | ||||||||||
| <30 | 1674 (57.4) | 1.45 | 652 | 66.5 | 525 | 56.3 | 383 | 51.1 | 114 | 44.7 | |
| >=30 | 1238 (42.4) | 2.14 | 328 | 33.4 | 406 | 43.5 | 363 | 48.4 | 141 | 55.3 | |
| Physical Activity (MET-h/wk) | <0.001 | ||||||||||
| Q1 (0–2) | 572 (19.6) | 2.75 | 109 | 11.1 | 167 | 17.9 | 217 | 28.9 | 79 | 31.0 | |
| Q2 (2.25–7.2) | 553 (18.9) | 2.24 | 137 | 14.0 | 186 | 19.9 | 160 | 21.3 | 70 | 27.5 | |
| Q3 (7.25–14) | 572 (19.6) | 1.65 | 205 | 20.9 | 185 | 19.8 | 139 | 18.5 | 46 | 18.0 | |
| Q4 (14.1–24) | 560 (19.2) | 1.35 | 232 | 23.6 | 179 | 19.2 | 121 | 16.1 | 28 | 11.0 | |
| Q5 (24.2–107) | 565 (19.4) | 1.07 | 263 | 26.8 | 187 | 20.0 | 90 | 12.0 | 22 | 8.6 | |
P-values were calculated using chi-square tests for categorical variables.
Table 2.
Serum concentrations of C-reactive Protein by tumor characteristics in breast cancer survivors of the Women's Healthy Eating and Living Study
| Distribution of Inflammatory Status | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| No Inflammation |
Low Inflammation |
Moderate Inflammation |
Acute Inflammation |
||||||||
| Serum CRP mg/L | <1 mg/L | 1–3 mg/L | >3 to <10 mg/L | ≥10 mg/L | |||||||
| n (%) | Geometric mean | No. | % | No. | % | No. | % | No. | % | P | |
| Tumor Stage | 0.007 | ||||||||||
| I | 1144 (39.2) | 1.54 (0.58, 3.91) | 418 | 42.6 | 360 | 38.6 | 277 | 36.9 | 89 | 34.9 | |
| II | 1333 (45.7) | 1.73 (0.71, 4.17) | 527 | 53.7 | 538 | 57.7 | 424 | 56.5 | 150 | 58.8 | |
| IIIA | 442 (15.1) | 2.2 (0.94, 5.55) | 36 | 3.7 | 35 | 3.8 | 49 | 6.5 | 16 | 6.3 | |
| Tumor Grade | 0.097 | ||||||||||
| 1 - Well differentiated | 463 (15.9) | 1.51 (0.54, 4.23) | 181 | 18.5 | 131 | 14.0 | 109 | 14.5 | 42 | 16.5 | |
| 2 - Moderately differentiated | 1184 (40.6) | 1.64 (0.66, 3.88) | 397 | 40.5 | 395 | 42.3 | 302 | 40.3 | 90 | 35.3 | |
| 3 - Poorly differentiated | 1027 (35.2) | 1.92 (0.77, 4.87) | 316 | 32.2 | 331 | 35.5 | 278 | 37.1 | 102 | 40.0 | |
| Unspecified | 245 (8.4) | 1.66 (0.63, 4.04) | 87 | 8.9 | 76 | 8.1 | 61 | 8.1 | 21 | 8.2 | |
| Estrogen/Progesterone Receptor Status | 0.003 | ||||||||||
| ER+/PR+ | 1818 (62.3) | 1.60 (0.64, 4.02) | 631 | 64.3 | 597 | 64.0 | 459 | 61.2 | 135 | 52.9 | |
| ER+/PR− | 331 (11.3) | 1.53 (0.63, 3.83) | 121 | 12.3 | 106 | 11.4 | 84 | 11.2 | 22 | 8.6 | |
| ER−/PR+ | 121 (4.2) | 1.73 (0.57, 4.35) | 42 | 4.3 | 33 | 3.5 | 34 | 4.5 | 12 | 4.7 | |
| ER−/PR− | 582 (19.9) | 2.20 (0.85, 5.50) | 170 | 17.3 | 179 | 19.2 | 154 | 20.5 | 78 | 30.6 | |
| Unknown | 67 (2.3) | 2.20 (0.91, 6.06) | 17 | 1.7 | 18 | 1.9 | 19 | 2.5 | 8 | 3.1 | |
| Her2 Receptor Status | 0.605 | ||||||||||
| Positive | 360 (12.3) | 1.72 (0.66, 4.22) | 123 | 12.5 | 108 | 11.6 | 100 | 13.3 | 29 | 11.4 | |
| Negative | 1672 (57.3) | 1.74 (0.70,4.26) | 542 | 55.2 | 543 | 58.2 | 438 | 58.4 | 149 | 58.4 | |
| Unknown | 887 (30.4) | 1.66 (0.65, 4.17) | 316 | 32.2 | 282 | 30.2 | 212 | 28.3 | 77 | 30.2 | |
| Treatment (after surgery) | 0.597 | ||||||||||
| No chemotherapy or radiation | 321 (11.0) | 1.56 (0.58, 3.86) | 112 | 11.4 | 106 | 11.4 | 74 | 9.9 | 29 | 11.4 | |
| Radiation only | 569 (19.5) | 1.63 (0.66, 4.00) | 196 | 20.0 | 193 | 20.7 | 140 | 18.7 | 40 | 15.7 | |
| Chemotherapy only | 794 (27.2) | 1.76 (0.66, 4.17) | 255 | 26.0 | 258 | 27.7 | 205 | 27.3 | 76 | 29.8 | |
| Radiation and chemotherapy | 1230 (42.1) | 1.76 (0.71, 4.56) | 415 | 42.3 | 375 | 40.2 | 331 | 44.1 | 109 | 42.7 | |
| Unknown | 5 (0.17) | 0.85 (0.25,2.68) | 3 | 0.3 | 1 | 0.1 | 0 | 0.0 | 1 | 0.4 | |
| Current Anti-estrogen use | <0.001 | ||||||||||
| Yes | 1796 (61.5) | 1.54 (0.63, 3.92) | 640 | 65.2 | 586 | 62.8 | 449 | 59.9 | 121 | 47.5 | |
| No | 1119 (38.3) | 2.03 (0.80, 5.10) | 341 | 34.8 | 343 | 36.8 | 301 | 40.1 | 134 | 52.5 | |
| Unknown | 4 (0.13) | 1.75 (1.29, 2.37) | 0 | 0.0 | 4 | 0.4 | 0 | 0.0 | 0 | 0.0 | |
P-values were calculated using chi-square tests for categorical variables.
A total of 236 women died including 188 deaths due to breast cancer, and 417 women experienced an additional breast cancer event (recurrence [85%] or new breast cancer primary [15%]) during a median follow-up period of 7.4 years. Higher vs. lower concentrations of CRP were associated with increased risk for all-cause mortality, breast cancer-specific mortality, and additional breast cancer events (Ptrend <0.05 in all models; lnCRP, P <0.05 in all models) (Table 3). The 5-year unadjusted overall survival rates in women with acute, moderate, low and no inflammation were 96.1%, 98.4%, 98.4%, and 97.8 %, respectively (log rank p<0.001, data not shown). The comparable 10-year figures were 85.8%, 89.8%, 92.6%, and 93.0%, respectively (log rank p=0.002, data not shown).
Table 3.
Risk of all-cause mortality, breast cancer-specific mortality, and additional breast cancer event by inflammatory status using serum CRP concentrations in breast cancer survivors in the Women's Healthy Eating and Living Study
| Model 1a | Model 2b | Model 3c | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Models of association by serum CRP |
No. of Events |
N | HR | 95% CI | Ptrend | HR | 95% CI | Ptrend | HR | 95% CI | Ptrend |
| All-cause Mortality | |||||||||||
| Inflammatory Status | <0.001 | 0.005 | 0.006 | ||||||||
| No inflammation, CRP <1 mg/L | 61 | 981 | 1.00 | referrent | 1.00 | referrent | 1.00 | referrent | |||
| Low, CRP 1 to 3 mg/L | 66 | 933 | 1.01 | 0.71–1.44 | 0.99 | 0.69–1.43 | 0.98 | 0.68–1.42 | |||
| Moderate, CRP >3 to <10 mg/L | 73 | 750 | 1.41 | 0.99–2.00 | 1.32 | 0.89–1.96 | 1.30 | 0.88–1.93 | |||
| Acute, CRP ≥10 mg/L | 36 | 255 | 2.12 | 1.38–3.23 | 1.96 | 1.22–3.13 | 1.92 | 1.20–3.08 | |||
| lnCRP | 236 | 2919 | 1.21 | 1.09–1.34 | 1.19 | 1.06–1.34 | 1.19 | 1.05–1.34 | |||
| Breast cancer Mortality | |||||||||||
| Inflammatory Status | 0.004 | 0.03 | 0.03 | ||||||||
| No inflammation | 51 | 981 | 1.00 | referent | 1.00 | referrent | 1.00 | referrent | |||
| Low Inflammation | 56 | 933 | 1.10 | 0.75–1.62 | 1.07 | 0.71–1.59 | 1.07 | 0.72–1.59 | |||
| Moderate Inflammation | 53 | 750 | 1.33 | 0.90–1.97 | 1.22 | 0.78–1.90 | 1.22 | 0.78–1.91 | |||
| Acute Inflammation | 28 | 255 | 2.10 | 1.30–3.38 | 1.91 | 1.13–3.23 | 1.88 | 1.11–3.18 | |||
| lnCRP | 188 | 2919 | 1.18 | 1.06–1.33 | 1.16 | 1.02–1.32 | 1.16 | 1.01–1.31 | |||
| Additional breast cancer events | |||||||||||
| Inflammatory Status | 0.01 | 0.02 | 0.03 | ||||||||
| No inflammation | 126 | 981 | 1.00 | referrent | 1.00 | referrent | 1.00 | referrent | |||
| Low Inflammation | 126 | 933 | 1.06 | 0.82–1.36 | 1.08 | 0.82–1.40 | 1.06 | 0.82–1.38 | |||
| Moderate Inflammation | 111 | 750 | 1.12 | 0.86–1.45 | 1.14 | 0.85–1.53 | 1.12 | 0.83–1.50 | |||
| Acute Inflammation | 54 | 255 | 1.67 | 1.20–2.31 | 1.69 | 1.17–2.43 | 1.65 | 1.15–2.38 | |||
| lnCRP | 417 | 2919 | 1.12 | 1.04–1.21 | 1.14 | 1.04–1.24 | 1.13 | 1.03–1.24 | |||
Abbreviation: HR, hazard ratio; 95% CI, 95 percent confidence interval, CRP, C-reactive protein; BMI, body mass index
Adjusted for age (continuous), time since dx, disease stage, disease grade, and race-ethnicity
Adjusted for age (continuous), time since dx, disease stage, disease grade, race-ethnicity, and BMI (categorical)
Adjusted for age (continuous), time since dx, disease stage, disease grade, race-ethnicity, BMI (categorical), anti-estrogen use and estrogen receptor/progesterone receptor status
A one-unit increase in lnCRP, which corresponds to an approximately 2.7-fold increase in CRP level, was associated with a 21% increased risk of all-cause mortality (model 1) and an 18% increased risk of breast cancer-specific mortality (model 1). For example, a lnCRP increase from 0 to 1 translates to a serum CRP concentration increase from 1 mg/L to 2.7 mg/L and an associated HR for all-cause mortality of 1.21 and HR for breast cancer-specific mortality of 1.18 (Table 3); a lnCRP value of 2, equates to a serum CRP concentration of 7.4 mg/L with respective HRs of 1.46 and 1.39; and a lnCRP value of 3, is a serum CRP concentration >20 mg/L with HRs of 1.77 and 1.64, respectively, compared with CRP level of 1 mg/L. When CRP was modeled using four categories of inflammatory status, we saw a threshold effect. Compared to women with no inflammation (<1 mg/L), those with values in the acute range (≥10 mg/L) had an approximate 2-fold increased risks of mortality due to breast cancer and/or any cause. Women with moderate inflammation (>3 to <10 mg/L) had an increased risk of mortality due to any cause or breast cancer that ranged between 33 to 41%, although these associations were not statistically significant. Women with no inflammation and low inflammation (1 to 3 mg/L) had comparable risks for all-cause and breast cancer-specific mortality. Adjusting for BMI (model 2) plus anti-estrogen use and tumor hormone receptor status (model 3) did not meaningfully alter the CRP-survival relationship (Table 3).
Women with acute inflammation (vs. no inflammation) had a 67% increase risk of an additional breast cancer-related event, whereas, women with low to moderate inflammation (vs. no inflammation) had comparable HRs for an additional breast cancer-related event. Acute inflammatory status remained an independent predictor following adjustment for BMI (model 2) plus anti-estrogen use and tumor hormone status (model 3). In a sensitivity analysis, we limited the outcomes to breast cancer recurrences, and the findings were not meaningfully changed (data not shown).
We also examined the association between very high CRP (≥10 mg/L) and breast cancer-specific mortality by select subgroups. The adverse association between very high CRP and death due to breast cancer was stronger in certain subgroups (Table 4). These subgroups included older women (≥55 years of age), postmenopausal women, those with excessive adiposity (BMI ≥25 kg/m2), and women whose tumors were estrogen receptor positive. However, interaction terms for CRP with these factors were not statistically significant. Further, separate models examined the CRP-survival relationship stratified by 1) intervention arm and 2) time interval from diagnosis to blood draw; the hazard estimates for CRP did not vary by intervention arm or time interval value (data not shown).
Table 4.
Subgroup Analysis of elevated C-reactive Protein (>10 mg/L) and risk of breast cancer-specific mortality in breast cancer survivors in the Women's Healthy Eating and Living Study
| Stratified Multivariable- Adjusted Model |
|||||
|---|---|---|---|---|---|
| Stratified models of association by upper quartile of CRP |
No. of Events |
N | HR | 95% CI | Pa |
| Age, years | 0.92 | ||||
| <55 | 115 | 1936 | 1.42 | 0.81–2.50 | |
| ≥55 | 73 | 983 | 2.08 | 1.07–4.05 | |
| Postmenopausal | 0.74 | ||||
| No | 38 | 595 | 0.86 | 0.26–2.66 | |
| Yes | 150 | 2324 | 1.89 | 1.19–3.02 | |
| BMI, kg/m2 | 0.56 | ||||
| 18.5 to 24.9 | 70 | 1249 | 1.08 | 0.26–4.56 | |
| ≥25 | 119 | 1697 | 1.84 | 1.18–2.88 | |
| Tumor Stage | 0.32 | ||||
| I | 25 | 1144 | 2.51 | 0.82–7.64 | |
| II to III | 163 | 1775 | 1.55 | 0.97–2.47 | |
| Tumor Estrogen Receptor | 0.18 | ||||
| ER positive | 120 | 2171 | 1.98 | 1.17–3.38 | |
| ER negative | 66 | 710 | 1.33 | 0.65–2.72 | |
| Cardiovascular disease | 0.99 | ||||
| No history of CVD | 111 | 1801 | 1.44 | 0.93–2.22 | |
| History of CVD | 39 | 605 | 1.17 | 0.56–2.43 | |
| Arthritis | 0.64 | ||||
| No history of arthritis | 121 | 1944 | 1.47 | 0.94–2.59 | |
| History of arthritis | 29 | 462 | 2.82 | 0.99–8.00 | |
| Type 2 Diabetes | 0.65 | ||||
| No history of T2DM | 139 | 2295 | 1.46 | 0.86–2.50 | |
| History of T2DM | 11 | 111 | 2.27 | 0.55–9.44 | |
Abbreviation: HR, hazard ratio; 95% CI, 95 percent confidence interval, CRP, C-reactive protein; BMI, body mass index
Two-sided P for interaction
Adjusted for following variable, including age, time since diagnosis tumor, postmenopausal status, tumor stage, tumor grade, race-ethnicity, BMI, anti-estrogen use, and estrogen/progesterone receptor status, unless stratified by specific variable.
DISCUSSION
This analysis assessed the distribution and the prognostic value of serum CRP with all-cause and breast cancer-specific mortality and additional breast cancer-related events among a cohort of women following a diagnosis of breast cancer. In this cohort, we observed a prevalence (34.4%) of elevated CRP (>3 mg/L) that is similar to that of the general population (ranging 25 to 43%) (35). This concurrence suggests that circulating concentrations of CRP return to relatively normal ranges following cancer diagnosis and treatment. Compared to women with the lowest concentrations of CRP (<1 mg/L), those with values of CRP ≥10 mg/L had an approximate 2-fold increase risk of all-cause and breast cancer-specific mortality, and a 67% increased risk of an additional breast cancer-related events in models adjusted for prognostic confounders. We also found that acute inflammation-survival relationship was not modified by age, postmenopausal status, BMI, tumor stage, ER+ receptor tumor status, or inflammatory-related conditions.
These results, including the size of the HRs and statistical significance, are comparable to reports from two other observational prospective cohort studies of women following a diagnosis of breast cancer. The larger report was conducted in a cohort of 2910 Danish women diagnosed with invasive breast cancer (36). Women whose blood was drawn prior to treatment, Allin et al. reported women with CRP levels >3.24 mg/L (highest tertilevs. the lowest tertile, <1.04 mg/L) had increased risk of all-cause mortality (HR=1.84; 95%CI, 1.39–2.45), an increased risk of death due to breast cancer (HR=1.66; 95%CI, 1.15–2.41), and a suggested increased risk of disease recurrence (HR=1.45; 95%CI, 0.93–2.26), independent of age, tumor characteristics (size, grade, ER/PR receptor status, HER2 status), lymph node status, presence of distant metastases, lifestyle factors (smoking, alcohol consumption), BMI and CVD (18). In another cohort of 734 U.S. postmenopausal breast cancer survivors (stage I-IIIA) whose blood was drawn following disease treatment, Pierce et al. reported women in the highest tertile (≥3.9 mg/L) of CRP concentrations (vs. lowest tertile, ≤1.2 mg/L) had 2-fold increase risk of all-cause mortality (HR=2.31; 95%CI, 1.30–4.12) and disease-free survival (HR=1.99; 95%CI, 1.09–3.65) after adjusting for age, disease stage, race/study site, BMI, and ER/PR receptor status (19). Conversely, results from smaller studies (23–25) suggest no association between CRP and prognosis. The differences in HRs and statistical significance may be due to confounding, timing of blood draw with respect to disease treatment, or may suggest that these studies were inadequately powered to evaluate the association in the presence of potential confounding.
The role of CRP as a biological indicator of prognosis versus as a contributor to carcinogenesis is unclear. CRP has some immune-related functions, including activation of classical complement binding and opsonization (for phagocytosis) (37). In a recent study of genetic variants in the CRP gene, certain variants were associated with altered plasma CRP concentrations, but not cancer risk (2). In this context, our finding that very high concentrations of circulating CRP (>10 mg/L) was associated with increased risk of mortality due to breast cancer and additional breast cancer events suggest that CRP may be a biological indicator of carcinogenesis, rather than a direct contributor.
Strengths of this analysis include the cohort study design, almost complete outcomes follow-up conducted with medical and death records, a reasonably large sample size and relatively long follow-up time. Furthermore, the sample size of this cohort was sufficient to assess the effect of acute inflammation (>10 mg/L) and outcomes in breast cancer survivors, in contrast to the other papers cited where 'Moderate' and 'Acute' inflammation (as defined here) were combined. The inclusion of the acute inflammation category leads to the suggestion of a possible threshold effect. Thus, acute inflammation status (CRP ≥10 mg/L) may be an important independent biomarker for long-term survival in breast cancer survivors. A limitation of our study was that our samples were obtained at one time-point following completion of active treatment (on average 2 years following diagnosis); consequently no conclusions can be drawn as to whether elevated CRP persisted over the long term. Future studies using blood samples from multiple time-points before and after cancer diagnosis may provide important information clarifying the temporal relationship of biomarkers with cancer survival. Lastly, our results are only generalizable to women who have completed treatment and survived at least 2 years after diagnosis of breast cancer.
In summary, our results indicate that the risk of mortality and disease recurrence among women diagnosed with breast cancer is associated with acute inflammation (>10 mg/L) and possibly moderate inflammation (CRP >3mg/L). Combined with previous research, this study appears to confirm the association of circulating CRP levels with increased risk of poor breast cancer outcomes. It is particularly notable that there was an average of 7 years from the baseline CRP assessment and recurrence or mortality, which indicates that CRP is not a transient measure of acute inflammation in this cohort but offers long-term prognostic information. Arguably, the most important next step is to determine whether this association is causal. If CRP plays a causal role in breast cancer outcomes, then future research should be focused on understanding how lifestyle interventions such as diet and physical activity can reduce the circulating concentrations of this risk factor. Alternatively, if CRP is simply a prognostic marker, then further research is needed to understand whether it is responsive (i.e., has utility) to drug and lifestyle interventions designed to decrease risk of recurrence.
Acknowledgements
The authors thank Jennifer A. Emond, Gail A. Laughlin, and the WHEL research team for their expertise and hard work on the study, in addition to all the study participants who shared their personal and medical information to make this research possible.
Funding/Support: This work was supported by funding from the Walton Family Foundation; National Cancer Institute grant CA 69375; National Institutes of Health grants M01-RR00070, M01-RR00079, and M01-RR00827; and the KOMEN Foundation (Grant 100988) to PI John P. Pierce. In addition, Catherine R. Marinac received support from Ruth L. Kirschstein National Research Service Award Institutional training grant (5 T32 GM084896).
Footnotes
Author Contributions: Drs. Villasenor, Pierce and Patterson and Ms. Flatt had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the analysis. Study concept and design: Villasenor, Pierce, Patterson; Acquisition of data: Pierce, Flatt; Analysis and interpretation of data: Villasenor, Flatt, Natarajan, and Patterson; Drafting of the manuscript: Villasenor; Critical revision of the manuscript for important intellectual content: Pierce, Patterson, Natarajan, Flatt, Marinac; Statistical analysis: Flatt, Villasenor; Obtained funding: Pierce; Administrative, technical, or material support: Flatt; Study supervision: Patterson, Pierce.
Conflict of Interest: The authors have no disclosures.
REFERENCES
- 1.Society AC. Breast Cancer Facts & Figures 2011–2012. Atlanta, GA: American Cancer Society, Inc.; 2012. [Google Scholar]
- 2.Allin KH, Nordestgaard BG, Zacho J, Tybjaerg-Hansen A, Bojesen SE. C-reactive protein and the risk of cancer: a Mendelian randomization study. J Natl Cancer Inst. 2009;102(3):202–206. doi: 10.1093/jnci/djp459. [DOI] [PubMed] [Google Scholar]
- 3.Mantovani A, Allavena P, Sica A, Balkwill F. Cancer-related inflammation. Nature. 2008 Jul 24;454(7203):436–444. doi: 10.1038/nature07205. PubMed PMID: 18650914. Epub 2008/07/25. eng. [DOI] [PubMed] [Google Scholar]
- 4.Sica A, Allavena P, Mantovani A. Cancer related inflammation: the macrophage connection. Cancer letters. 2008 Aug 28;267(2):204–215. doi: 10.1016/j.canlet.2008.03.028. PubMed PMID: 18448242. Epub 2008/05/02. eng. [DOI] [PubMed] [Google Scholar]
- 5.Chiang AC, Massague J. Molecular basis of metastasis. The New England journal of medicine. 2008 Dec 25;359(26):2814–2823. doi: 10.1056/NEJMra0805239. PubMed PMID: 19109576. Epub 2008/12/26. eng. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Favaro E, Amadori A, Indraccolo S. Cellular interactions in the vascular niche: implications in the regulation of tumor dormancy. APMIS : acta pathologica, microbiologica, et immunologica Scandinavica. 2008 Jul-Aug;116(7–8):648–659. doi: 10.1111/j.1600-0463.2008.01025.x. PubMed PMID: 18834409. Epub 2008/10/07. eng. [DOI] [PubMed] [Google Scholar]
- 7.Seruga B, Zhang H, Bernstein LJ, Tannock IF. Cytokines and their relationship to the symptoms and outcome of cancer. Nature reviews Cancer. 2008 Nov;8(11):887–899. doi: 10.1038/nrc2507. PubMed PMID: 18846100. Epub 2008/10/11. eng. [DOI] [PubMed] [Google Scholar]
- 8.Ikeda M, Natsugoe S, Ueno S, Baba M, Aikou T. Significant host- and tumor-related factors for predicting prognosis in patients with esophageal carcinoma. Annals of surgery. 2003 Aug;238(2):197–202. doi: 10.1097/01.sla.0000080822.22415.cb. PubMed PMID: 12894012. Pubmed Central PMCID: 1422696. Epub 2003/08/02. eng. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.McMillan DC, Canna K, McArdle CS. Systemic inflammatory response predicts survival following curative resection of colorectal cancer. The British journal of surgery. 2003 Feb;90(2):215–219. doi: 10.1002/bjs.4038. PubMed PMID: 12555298. Epub 2003/01/30. eng. [DOI] [PubMed] [Google Scholar]
- 10.Hilmy M, Bartlett JM, Underwood MA, McMillan DC. The relationship between the systemic inflammatory response and survival in patients with transitional cell carcinoma of the urinary bladder. British journal of cancer. 2005 Feb 28;92(4):625–627. doi: 10.1038/sj.bjc.6602406. PubMed PMID: 15726119. Pubmed Central PMCID: 2361870. Epub 2005/02/24. eng. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Jamieson NB, Glen P, McMillan DC, McKay CJ, Foulis AK, Carter R, et al. Systemic inflammatory response predicts outcome in patients undergoing resection for ductal adenocarcinoma head of pancreas. British journal of cancer. 2005 Jan 17;92(1):21–23. doi: 10.1038/sj.bjc.6602305. PubMed PMID: 15597096. Pubmed Central PMCID: 2361749. Epub 2004/12/15. eng. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Roxburgh CS, McMillan DC. Role of systemic inflammatory response in predicting survival in patients with primary operable cancer. Future oncology. 2010 Jan;6(1):149–163. doi: 10.2217/fon.09.136. PubMed PMID: 20021215. Epub 2009/12/22. eng. [DOI] [PubMed] [Google Scholar]
- 13.Schmid M, Schneitter A, Hinterberger S, Seeber J, Reinthaller A, Hefler L. Association of elevated C-reactive protein levels with an impaired prognosis in patients with surgically treated endometrial cancer. Obstetrics and gynecology. 2007 Dec;110(6):1231–1236. doi: 10.1097/01.AOG.0000292085.50987.f2. PubMed PMID: 18055714. Epub 2007/12/07. eng. [DOI] [PubMed] [Google Scholar]
- 14.Polterauer S, Grimm C, Tempfer C, Sliutz G, Speiser P, Reinthaller A, et al. C-reactive protein is a prognostic parameter in patients with cervical cancer. Gynecologic oncology. 2007 Oct;107(1):114–117. doi: 10.1016/j.ygyno.2007.06.001. PubMed PMID: 17617445. Epub 2007/07/10. eng. [DOI] [PubMed] [Google Scholar]
- 15.O'Dowd C, McRae LA, McMillan DC, Kirk A, Milroy R. Elevated preoperative C-reactive protein predicts poor cancer specific survival in patients undergoing resection for non-small cell lung cancer. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer. 2010 Jul;5(7):988–992. doi: 10.1097/JTO.0b013e3181da78f9. PubMed PMID: 20453690. Epub 2010/05/11. eng. [DOI] [PubMed] [Google Scholar]
- 16.Allin KH, Nordestgaard BG. Elevated C-reactive protein in the diagnosis, prognosis, and cause of cancer. Critical reviews in clinical laboratory sciences. 2011 Jul-Aug;48(4):155–170. doi: 10.3109/10408363.2011.599831. PubMed PMID: 22035340. Epub 2011/11/01. eng. [DOI] [PubMed] [Google Scholar]
- 17.Pearson TA, Mensah GA, Alexander RW, Anderson JL, Cannon RO, 3rd, Criqui M, et al. Markers of inflammation and cardiovascular disease: application to clinical and public health practice: A statement for healthcare professionals from the Centers for Disease Control and Prevention and the American Heart Association. Circulation. 2003 Jan 28;107(3):499–511. doi: 10.1161/01.cir.0000052939.59093.45. PubMed PMID: 12551878. Epub 2003/01/29. eng. [DOI] [PubMed] [Google Scholar]
- 18.Allin KH, Nordestgaard BG, Flyger H, Bojesen SE. Elevated pre-treatment levels of plasma C-reactive protein are associated with poor prognosis after breast cancer: a cohort study. Breast cancer research : BCR. 2011;13(3):R55. doi: 10.1186/bcr2891. PubMed PMID: 21639875. Pubmed Central PMCID: 3218944. Epub 2011/06/07. eng. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Pierce BL, Ballard-Barbash R, Bernstein L, Baumgartner RN, Neuhouser ML, Wener MH, et al. Elevated biomarkers of inflammation are associated with reduced survival among breast cancer patients. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2009 Jul 20;27(21):3437–3444. doi: 10.1200/JCO.2008.18.9068. PubMed PMID: 19470939. Pubmed Central PMCID: 2717751. Epub 2009/05/28. eng. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Al Murri AM, Bartlett JM, Canney PA, Doughty JC, Wilson C, McMillan DC. Evaluation of an inflammation-based prognostic score (GPS) in patients with metastatic breast cancer. British journal of cancer. 2006 Jan 30;94(2):227–230. doi: 10.1038/sj.bjc.6602922. PubMed PMID: 16404432. Pubmed Central PMCID: 2361117. Epub 2006/01/13. eng. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Albuquerque KV, Price MR, Badley RA, Jonrup I, Pearson D, Blamey RW, et al. Pre-treatment serum levels of tumour markers in metastatic breast cancer: a prospective assessment of their role in predicting response to therapy and survival. European journal of surgical oncology : the journal of the European Society of Surgical Oncology and the British Association of Surgical Oncology. 1995 Oct;21(5):504–509. doi: 10.1016/s0748-7983(95)96935-7. PubMed PMID: 7589594. Epub 1995/10/01. eng. [DOI] [PubMed] [Google Scholar]
- 22.Williams MR, Turkes A, Pearson D, Griffiths K, Blamey RW. An objective biochemical assessment of therapeutic response in metastatic breast cancer: a study with external review of clinical data. British journal of cancer. 1990 Jan;61(1):126–132. doi: 10.1038/bjc.1990.26. PubMed PMID: 2137007. Pubmed Central PMCID: 1971340. Epub 1990/01/01. eng. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Al Murri AM, Wilson C, Lannigan A, Doughty JC, Angerson WJ, McArdle CS, et al. Evaluation of the relationship between the systemic inflammatory response and cancer-specific survival in patients with primary operable breast cancer. British journal of cancer. 2007 Mar 26;96(6):891–895. doi: 10.1038/sj.bjc.6603682. PubMed PMID: 17375036. Pubmed Central PMCID: 2360103. Epub 2007/03/22. eng. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Pasanisi P, Venturelli E, Morelli D, Fontana L, Secreto G, Berrino F. Serum insulin-like growth factor-I and platelet-derived growth factor as biomarkers of breast cancer prognosis. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology. 2008 Jul;17(7):1719–1722. doi: 10.1158/1055-9965.EPI-07-0654. PubMed PMID: 18628423. Epub 2008/07/17. eng. [DOI] [PubMed] [Google Scholar]
- 25.Ravishankaran P, Karunanithi R. Clinical significance of preoperative serum interleukin-6 and C-reactive protein level in breast cancer patients. World journal of surgical oncology. 2011;9:18. doi: 10.1186/1477-7819-9-18. PubMed PMID: 21294915. Pubmed Central PMCID: 3045973. Epub 2011/02/08. eng. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Han Y, Mao F, Wu Y, Fu X, Zhu X, Zhou S, et al. Prognostic role of C-reactive protein in breast cancer: a systematic review and meta-analysis. The International journal of biological markers. 2011 Oct-Dec;26(4):209–215. doi: 10.5301/JBM.2011.8872. PubMed PMID: 22139643. Epub 2011/12/06. eng. [DOI] [PubMed] [Google Scholar]
- 27.Pierce JP, Faerber S, Wright FA, Rock CL, Newman V, Flatt SW, et al. A randomized trial of the effect of a plant-based dietary pattern on additional breast cancer events and survival: the Women's Healthy Eating and Living (WHEL) Study. Controlled clinical trials. 2002 Dec;23(6):728–756. doi: 10.1016/s0197-2456(02)00241-6. PubMed PMID: 12505249. Epub 2002/12/31. eng. [DOI] [PubMed] [Google Scholar]
- 28.Pierce JP, Natarajan L, Caan BJ, Parker BA, Greenberg ER, Flatt SW, et al. Influence of a diet very high in vegetables, fruit, and fiber and low in fat on prognosis following treatment for breast cancer: the Women's Healthy Eating and Living (WHEL) randomized trial. AMA : the journal of the American Medical Association. 2007 Jul 18;298(3):289–298. doi: 10.1001/jama.298.3.289. PubMed PMID: 17635889. Pubmed Central PMCID: 2083253. Epub 2007/07/20. eng. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Johnson-Kozlow M, Rock CL, Gilpin EA, Hollenbach KA, Pierce JP. Validation of the WHI brief physical activity questionnaire among women diagnosed with breast cancer. American journal of health behavior. 2007 Mar-Apr;31(2):193–202. doi: 10.5555/ajhb.2007.31.2.193. PubMed PMID: 17269909. Epub 2007/02/03. eng. [DOI] [PubMed] [Google Scholar]
- 30.Thomson CA, Giuliano A, Rock CL, Ritenbaugh CK, Flatt SW, Faerber S, et al. Measuring dietary change in a diet intervention trial: comparing food frequency questionnaire and dietary recalls. American journal of epidemiology. 2003 Apr 15;157(8):754–762. doi: 10.1093/aje/kwg025. PubMed PMID: 12697580. Epub 2003/04/17. eng. [DOI] [PubMed] [Google Scholar]
- 31.Hong S, Bardwell WA, Natarajan L, Flatt SW, Rock CL, Newman VA, et al. Correlates of physical activity level in breast cancer survivors participating in the Women's Healthy Eating and Living (WHEL) Study. Breast cancer research and treatment. 2007 Jan;101(2):225–232. doi: 10.1007/s10549-006-9284-y. PubMed PMID: 17028988. Epub 2006/10/10. eng. [DOI] [PubMed] [Google Scholar]
- 32.Kleinbaum DG, Klein M. Evaluating the proportional hazards assumption. New York: Springer; 2005. pp. 131–171. [Google Scholar]
- 33.Nguyen XM, Lane J, Smith BR, Nguyen NT. Changes in inflammatory biomarkers across weight classes in a representative US population: a link between obesity and inflammation. Journal of gastrointestinal surgery : official journal of the Society for Surgery of the Alimentary Tract. 2009 Jul;13(7):1205–1212. doi: 10.1007/s11605-009-0904-9. PubMed PMID: 19415399. Pubmed Central PMCID: 2693771. Epub 2009/05/06. eng. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Harrell FE, Jr, Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Statistics in medicine. 1996 Feb 28;15(4):361–387. doi: 10.1002/(SICI)1097-0258(19960229)15:4<361::AID-SIM168>3.0.CO;2-4. PubMed PMID: 8668867. Epub 1996/02/28. eng. [DOI] [PubMed] [Google Scholar]
- 35.Ford ES, Giles WH, Mokdad AH, Myers GL. Distribution and correlates of C-reactive protein concentrations among adult US women. Clinical chemistry. 2004 Mar;50(3):574–581. doi: 10.1373/clinchem.2003.027359. PubMed PMID: 14709450. Epub 2004/01/08. eng. [DOI] [PubMed] [Google Scholar]
- 36.Allin KH, Bojesen SE, Nordestgaard BG. Baseline C-reactive protein is associated with incident cancer and survival in patients with cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2009 May 1;27(13):2217–2224. doi: 10.1200/JCO.2008.19.8440. PubMed PMID: 19289618. Epub 2009/03/18. eng. [DOI] [PubMed] [Google Scholar]
- 37.Schultz DR, Arnold PI. Properties of four acute phase proteins: C-reactive protein, serum amyloid A protein, alpha 1-acid glycoprotein, and fibrinogen. Seminars in arthritis and rheumatism. 1990 Dec;20(3):129–147. doi: 10.1016/0049-0172(90)90055-k. PubMed PMID: 1705051. Epub 1990/12/01. eng. [DOI] [PubMed] [Google Scholar]