Background:

Previous studies have found that acute febrile infection may decrease the risk of breast cancer. Meanwhile, it is well known that interleukin-6 (IL6) played dual roles in the tumor microenvironment. Fever may stimulate IL6 production, and IL6 rs1800796 also influences the expression of IL6. However, the impact of fever and its interaction with IL6 rs1800796 on breast cancer survival remains to be explored.

Methods:

This was a prospective cohort study of 4,223 breast cancer patients. Exposures were pre-/postdiagnostic infection-induced fever and rs1800796 polymorphism. The endpoints were overall survival (OS) and progression-free survival (PFS). Adjusted hazard ratios were obtained using multivariate Cox proportional hazards regression models.

Results:

Compared with women without prediagnostic fever, the adjusted hazard ratio (HR) of progression for those with prediagnostic fever was 0.81 (95% CI, 0.66–0.99), particularly for the CC genotype of IL6 rs1800796 (HR, 0.53; 95% CI, 0.36–0.79). OS was also better (HR, 0.59; 95% CI, 0.36–0.99) among women with the CC genotype exposed to prediagnostic fever, accompanied by a significant interaction (P = 0.021). Postdiagnostic fever conferred better PFS for breast cancer (HR, 0.72; 95% CI, 0.52–1.00). Irrespective of the genotype of IL6, lymph node–positive women with postdiagnostic fever (HR, 0.57; 95% CI, 0.37–0.89) had a lower risk of progression than lymph node–negative women (HR, 1.12; 95% CI, 0.70–1.79).

Conclusions:

Infection-induced fever was beneficial to breast cancer survival, particularly for women who were the CC genotype of IL6 rs1800796 or node positive.

Impact:

This study provides new insight into the roles of infection-induced fever as a potential prognostic marker and therapy regimen for breast cancer.

Fever is an integrated physiologic and neuronal response as a hallmark of infection and inflammatory diseases (1). Numerous epidemiologic studies consistently found that exposures to acute febrile infection were associated with a reduced risk for many types of cancer (2). Since the 19th century, spontaneous cancer regression has been repeatedly observed in cancer patients with coincident infections (3). Inspired by this phenomenon, William Coley used heat-killed bacteria to treat cancer patients, known as “Coley's toxin.” Although some cancer patients experienced side effects, Coley's toxin cured some cancer patients (4, 5). Most recently, a case report showed that a patient with lymphoma experienced widespread resolution of lymphadenopathy after the infection of SARS-CoV-2 without any interventions (6). An ecologic study also found that malaria, a typical acute febrile infection, was associated with decreased mortality for some solid cancers, particularly for breast cancer (7).

Breast cancer is the most common cancer affecting women's health worldwide (8). There is a great need to identify predictors of breast cancer recurrence and mortality (9, 10). We have found a possible link between prediagnostic acute febrile infection and a decreased risk of breast cancer (11). It would be interesting and meaningful to explore the impact of the infection-induced fever on the prognosis. In addition, postdiagnostic fever is a common event for breast cancer patients, whereas only two small studies examined the effect of postdiagnostic fever on the prognosis (12, 13). Moreover, these two previous studies either limited the study subjects to node-negative patients or only evaluated the effect of fever during hospitalization without considering the impact of fever frequency after discharge.

Furthermore, it has been found that fever could stimulate the immune system by increasing the activity of T and natural killer (NK) cells and producing cytokines, such as IL6 (1, 14–16). IL6 is a pleiotropic cytokine with roles in innate and adaptive immunity (17) and has been implicated in both tumor progression and antitumor immunity (18). The IL6 expression has been found to increase with rising temperature in vivo and in vitro (19–21). In turn, as a crucial pyrogenic cytokine, IL6 is able to induce fever (22, 23). In addition, several functional single-nucleotide polymorphisms (SNP) of the IL6 gene promoter, providing genetic surrogates for IL6 production, have been reported to be associated with breast cancer survival (24, 25). However, evidence on the combined effects of functional IL6 promoter variants and fever on the survival of breast cancer patients is scarce.

Accordingly, we analyzed the associations of pre-/postdiagnostic infection-induced fever with the survival of breast cancer patients and the modification effects of IL6 rs1800796 polymorphism and the various subtypes, using data from the Guangzhou Breast Cancer Study (GZBCS) in China.

Study population

A total of 4,996 female patients, who were histologically diagnosed with primary breast cancer, were recruited from the First and Second Affiliated Hospitals and the Cancer Center of Sun Yat-sen University in Guangzhou, China, from October 2008 to January 2018. Those who missed prediagnostic infection-induced fever history (N = 298), lost follow-up (N = 317), were IV stage (N = 141), and self-reported history of other cancer (N = 17) were excluded, yielding an analytic sample of 4,223 cases. A written informed consent was obtained from all the participants. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethical Committee of the School of Public Health at Sun Yat-sen University (number of IRB approval:2012-8).

Data collection

Demographic information, including age, menopausal status, family history of cancer, body mass index (BMI), breastfeeding, and infection-induced fever history, were collected by trained investigators using the structural questionnaire as previously described (11). In the instance of infection-induced fever, patients were asked to recall the average times of fever per year over the past 10 years as the result of acute febrile infections, including influenza, common cold, abscess, bronchitis, pneumonia, and herpes simplex (no, <1, 1–2, 3–4, or ≥5 times). During the follow-up, patients were asked to recall the infection-induced fever frequency after diagnosis, using the same questions mentioned above. The interviewers did not know the hypotheses of this study. Clinicopathologic characteristics, including tumor size, lymph node status, and clinical stage, were collected from the medical record. Detailed definitions of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth receptor (HER2) status were previously described (26).

Follow-up

The patients were followed up at least every three months during the first year and every six months during the second and the third years; thereafter, patients were followed up annually until death or December 31, 2019. A total of 4,223 breast cancer patients were successfully obtained with a median follow-up time of 43.51 months. Follow-up information regarding recurrence, metastasis, and survival status was collected through phone calls, correspondence, and outpatient visits. The primary endpoint for this study was overall survival (OS), defined as the time from diagnosis until death; the patients still alive have been censored at their latest date of follow-up. The second endpoint was progression-free survival (PFS), calculated from diagnosis to the date of progression (recurrence or metastasis) or death; the patients still alive without progression have been censored at the latest follow-up date.

DNA isolation and genotyping

Blood samples of patients were collected immediately after admission to the hospital and were stored at −80°C until genotyping. Genomic DNA was extracted from the buffy coats of the participants using the TIANamp Genomic DNA kit (TianGen Biotech Co. Ltd) and genotyped using a matrix-assisted laser desorption/ionization time-of-flight mass spectrometry platform (Sequenom MassArray platform; Sequenom Inc.) under the manufacturer's instructions (27). Primers for the IL6 rs1800796 were previously described (28). Duplicate samples (5% of the total) were included to evaluate the genotyping quality, and the concordance rate was 100%. As our previous study described, IL6 rs1800796 genotyping was performed among 1,173 (27.8%) patients who were recruited from 2008 to 2012 (28). We didn't perform genotyping among the patients recruited after 2012 or without samples.

Statistical analysis

The potential confounders were determined by evaluating the associations of demographic and clinical characteristics with breast cancer OS and PFS. Then, Cox proportional hazards regression models were used to estimate the association between pre-/postdiagnostic infection-induced fever and the prognosis with hazard ratio (HR) and 95% CI, adjusting for age at diagnosis, education, menopausal status, ER and HER2 statuses, clinical stage, and Ki-67. We performed trend tests using categorical variables as continuous variables in the model to evaluate the dose–response relationship between the infection-induced fever frequency and breast cancer prognosis. We also evaluated the effect of the change of infection-induced fever status on breast cancer OS and PFS using the adjusted Cox proportional hazards model. Stratified analyses were further performed to assess whether the associations between pre-/postdiagnostic infection-induced fever and breast cancer survival were modified by IL6 rs1800796, lymph node status, and other clinical characteristics. Because of the low frequency of mutant homozygotes (GG), the mutant homozygotes (GG), and heterozygotes (CG) were combined for analysis in the dominant model. Interactions between these studied factors on the survival were estimated by the product terms in the Cox regression models. All statistical tests were two-tailed, and significance was assigned at P < 0.05. Statistical analyses above were conducted by R software (version 4.1.3).

Data availability

The data sets used and analyzed during this study are available from the corresponding author on reasonable request.

Demographic and clinicopathologic characteristics and the associations with breast cancer prognosis

As shown in Table 1, at the time of diagnosis, the mean age was 48.03 (SD = 10.60) years, and almost two thirds of them were premenopausal (61.4%). Nearly half of the participants had below junior school as their highest attained educational level (45.8%). Most women were ER-positive (71.7%), and almost three fourths of them were diagnosed with early cancer (stage I/II: 74.4%). Age at diagnosis, education status, menopausal status, ER status, PR status, tumor size, lymph nodes status, and clinical stage were significantly associated with breast cancer survival.

Table 1.

Demographic and clinicopathologic characteristics and the associations with breast cancer prognosis.

OSPFS
CharacteristicsTotal (%)EventsHR (95% CI)EventsHR (95% CI)
Age at diagnosis 
 ≤40 1,020 (24.2) 61 1.00 (reference) 165 1.00 (reference) 
 41–59 2,560 (60.6) 177 1.26 (0.94–1.68) 322 0.82 (0.68–0.99) 
 ≥60 641 (15.2) 58 1.61 (1.12–2.31) 108 1.09 (0.86–1.39) 
Education status 
 Below junior school 1,933 (45.8) 165 1.00 (reference) 292 1.00 (reference) 
 Senior high school 996 (23.6) 59 0.63 (0.47–0.85) 138 0.86 (0.70–1.05) 
 College or above 1,061 (25.1) 56 0.64 (0.47–0.87) 134 0.86 (0.70–1.06) 
Marital status 
 Married/living as married 3,842 (91.0) 261 1.00 (reference) 534 1.00 (reference) 
 Unmarrieda 302 (7.1) 29 1.22 (0.83–1.78) 51 1.08 (0.81–1.44) 
Age at menarche 
 ≤12 542 (12.8) 30 1.00 (reference) 68 1.00 (reference) 
 >12 3,598 (85.2) 257 1.28 (0.88–1.86) 513 1.12 (0.87–1.45) 
Menopausal status 
 Premenopausal 2,595 (61.4) 154 1.00 (reference) 345 1.00 (reference) 
 Postmenopausal 1,535 (36.3) 136 1.53 (1.22–1.93) 240 1.20 (1.02–1.42) 
Parity 
 0 230 (5.4) 15 1.00 (reference) 35 1.00 (reference) 
 ≥1 3,957 (93.7) 279 1.13 (0.67–1.89) 557 0.97 (0.69–1.37) 
Breastfeeding 
 Absent 352 (8.3) 21 1.00 (reference) 48 1.00 (reference) 
 Present 3,541 (83.9) 257 1.26 (0.80–1.96) 503 1.08 (0.80–1.45) 
Family history of breast cancer 
 No 3,920 (92.8) 281 1.00 (reference) 557 1.00 (reference) 
 Yes 235 (5.6) 0.53 (0.26–1.06) 25 0.81 (0.54–1.21) 
BMI (kg/m2) 
 <18.5 245 (5.8) 17 0.93 (0.57–1.53) 38 1.03 (0.74–1.44) 
 18.5–23.9 2,390 (56.6) 168 1.00 (reference) 337 1.00 (reference) 
 ≥24.0 1,437 (34.0) 102 1.03 (0.80–1.31) 203 1.01 (0.85–1.20) 
ER status 
 Negative 955 (22.6) 111 1.00 (reference) 195 1.00 (reference) 
 Positive 3,029 (71.7) 166 0.47 (0.37–0.59) 362 0.57 (0.48–0.68) 
PR status 
 Negative 1,319 (31.2) 137 1.00 (reference) 248 1.00 (reference) 
 Positive 2,659 (63.0) 140 0.47 (0.37–0.59) 307 0.56 (0.47–0.66) 
HER2 
 Negative 2,102 (49.8) 156 1.00 (reference) 310 1.00 (reference) 
 Positive/equivocal 1,759 (41.7) 117 1.23 (0.97–1.57) 232 1.17 (0.98–1.38) 
Triple-negative breast cancer 
 No 3,451 (81.7) 222 1.00 (reference) 454 1.00 (reference) 
 Yes 395 (9.4) 50 1.81 (1.33–2.46) 86 1.55 (1.23–1.95) 
Ki-67 
 <14%+ 926 (21.9) 34 1.00 (reference) 86 1.00 (reference) 
 ≥14%+ 2,768 (65.5) 205 2.32 (1.61,3.33) 416 1.84 (1.46–2.32) 
Tumor size (cm) 
 ≤2.0 1,639 (38.8) 55 1.00 (reference) 146 1.00 (reference) 
 >2.0 2,279 (54.0) 205 2.65 (1.97–3.57) 391 1.94 (1.61–2.35) 
Lymph nodes status 
 Negative 2,206 (52.2) 77 1.00 (reference) 191 1.00 (reference) 
 Positive 1,764 (41.8) 181 3.22 (2.46–4.20) 343 2.52 (2.11–3.00) 
Clinical stage 
 I 1,131 (26.8) 30 1.00 (reference) 81 1.00 (reference) 
 II 2012 (47.6) 101 1.93 (1.28–2.90) 233 1.67 (1.29–2.14) 
 III 739 (17.5) 119 7.01 (4.69–10.46) 205 4.64 (3.58–6.00) 
OSPFS
CharacteristicsTotal (%)EventsHR (95% CI)EventsHR (95% CI)
Age at diagnosis 
 ≤40 1,020 (24.2) 61 1.00 (reference) 165 1.00 (reference) 
 41–59 2,560 (60.6) 177 1.26 (0.94–1.68) 322 0.82 (0.68–0.99) 
 ≥60 641 (15.2) 58 1.61 (1.12–2.31) 108 1.09 (0.86–1.39) 
Education status 
 Below junior school 1,933 (45.8) 165 1.00 (reference) 292 1.00 (reference) 
 Senior high school 996 (23.6) 59 0.63 (0.47–0.85) 138 0.86 (0.70–1.05) 
 College or above 1,061 (25.1) 56 0.64 (0.47–0.87) 134 0.86 (0.70–1.06) 
Marital status 
 Married/living as married 3,842 (91.0) 261 1.00 (reference) 534 1.00 (reference) 
 Unmarrieda 302 (7.1) 29 1.22 (0.83–1.78) 51 1.08 (0.81–1.44) 
Age at menarche 
 ≤12 542 (12.8) 30 1.00 (reference) 68 1.00 (reference) 
 >12 3,598 (85.2) 257 1.28 (0.88–1.86) 513 1.12 (0.87–1.45) 
Menopausal status 
 Premenopausal 2,595 (61.4) 154 1.00 (reference) 345 1.00 (reference) 
 Postmenopausal 1,535 (36.3) 136 1.53 (1.22–1.93) 240 1.20 (1.02–1.42) 
Parity 
 0 230 (5.4) 15 1.00 (reference) 35 1.00 (reference) 
 ≥1 3,957 (93.7) 279 1.13 (0.67–1.89) 557 0.97 (0.69–1.37) 
Breastfeeding 
 Absent 352 (8.3) 21 1.00 (reference) 48 1.00 (reference) 
 Present 3,541 (83.9) 257 1.26 (0.80–1.96) 503 1.08 (0.80–1.45) 
Family history of breast cancer 
 No 3,920 (92.8) 281 1.00 (reference) 557 1.00 (reference) 
 Yes 235 (5.6) 0.53 (0.26–1.06) 25 0.81 (0.54–1.21) 
BMI (kg/m2) 
 <18.5 245 (5.8) 17 0.93 (0.57–1.53) 38 1.03 (0.74–1.44) 
 18.5–23.9 2,390 (56.6) 168 1.00 (reference) 337 1.00 (reference) 
 ≥24.0 1,437 (34.0) 102 1.03 (0.80–1.31) 203 1.01 (0.85–1.20) 
ER status 
 Negative 955 (22.6) 111 1.00 (reference) 195 1.00 (reference) 
 Positive 3,029 (71.7) 166 0.47 (0.37–0.59) 362 0.57 (0.48–0.68) 
PR status 
 Negative 1,319 (31.2) 137 1.00 (reference) 248 1.00 (reference) 
 Positive 2,659 (63.0) 140 0.47 (0.37–0.59) 307 0.56 (0.47–0.66) 
HER2 
 Negative 2,102 (49.8) 156 1.00 (reference) 310 1.00 (reference) 
 Positive/equivocal 1,759 (41.7) 117 1.23 (0.97–1.57) 232 1.17 (0.98–1.38) 
Triple-negative breast cancer 
 No 3,451 (81.7) 222 1.00 (reference) 454 1.00 (reference) 
 Yes 395 (9.4) 50 1.81 (1.33–2.46) 86 1.55 (1.23–1.95) 
Ki-67 
 <14%+ 926 (21.9) 34 1.00 (reference) 86 1.00 (reference) 
 ≥14%+ 2,768 (65.5) 205 2.32 (1.61,3.33) 416 1.84 (1.46–2.32) 
Tumor size (cm) 
 ≤2.0 1,639 (38.8) 55 1.00 (reference) 146 1.00 (reference) 
 >2.0 2,279 (54.0) 205 2.65 (1.97–3.57) 391 1.94 (1.61–2.35) 
Lymph nodes status 
 Negative 2,206 (52.2) 77 1.00 (reference) 191 1.00 (reference) 
 Positive 1,764 (41.8) 181 3.22 (2.46–4.20) 343 2.52 (2.11–3.00) 
Clinical stage 
 I 1,131 (26.8) 30 1.00 (reference) 81 1.00 (reference) 
 II 2012 (47.6) 101 1.93 (1.28–2.90) 233 1.67 (1.29–2.14) 
 III 739 (17.5) 119 7.01 (4.69–10.46) 205 4.64 (3.58–6.00) 

Abbreviations: BMI, body mass index; CI, confidence interval; HR, hazard ratio.

aUnmarried status included never married, separated, divorced, and widowed.

Association of prediagnostic infection-induced fever with breast cancer survival

Although 53.0% of women had no infection-induced fever history before diagnosis, the infection-induced fever frequencies for the other patients were as follows: below once, 35.9%; 1–2 times, 9.0%; 3–4 times, 1.4%; ≥ 5 times, 0.7%. After adjustment for potential prognostic factors of breast cancer, women who had experienced more than 0 times but less than or equal to 2 times of fever per year (0 < n ≤ 2 times/year) exhibited a significantly decreased risk of progression compared with those without fever [HR (95% CI) = 0.81 (0.66, 0.99)], whereas those with>2 times of fever per year exhibited a nonsignificantly decreased risk of progression (HR = 0.77; 95% CI, 0.34–1.73; Table 2). The nonsignificant decrease in progression risk was probably the result of the small sample size at the level of >2 times of fever (only 87 cases). Given that the direction of the two HRs was the same, we combined these two levels, yielding a significantly decreased risk of breast cancer progression for prediagnostic infection-induced fever with HR and 95% CI of 0.81 (0.66–0.99) in the adjusted multivariate model (Table 2), whereas there was a nonsignificant association between prediagnostic infection-induced fever and OS (HR 0.81; 95% CI, 0.61–1.09; Table 2).

Table 2.

Univariate and multivariate COX regression for pre-/postdiagnostic infection-induced fever with breast cancer prognosis.

OSPFS
Fever frequency (n times/yr)TotalEventsHR (95% CI)aHR (95% CI)bEventsHR (95% CI)aHR (95% CI)b
Prediagnosis 
n = 0 2,240 143 1.00 (reference) 1.00 (reference) 298 1.00 (reference) 1.00 (reference) 
 0 < n ≤ 2 times/yr 1,896 148 0.89 (0.70–1.12) 0.82 (0.61–1.10) 286 0.87 (0.74–1.03) 0.81 (0.66–0.99) 
n > 2 times/yr 87 0.73 (0.30–1.77) 0.59 (0.15–2.40) 12 0.88 (0.50–1.58) 0.77 (0.34–1.73) 
P for trend   0.249 0.143  0.115 0.041 
n > 0 1983 153 0.88 (0.70–1.11) 0.81 (0.61–1.09) 298 0.87 (0.74–1.03) 0.81 (0.66–0.99) 
Postdiagnosis 
n = 0 2,056 75 1.00 (reference) 1.00 (reference) 237 1.00 (reference) 1.00 (reference) 
 0 < n ≤ 2 times/yr 445 19 0.85 (0.51–1.41) 0.71 (0.38–1.30) 60 0.94 (0.71–1.26) 0.72 (0.51–1.01) 
n > 2 times/yr 56 0.30 (0.04–2.13) 0.34 (0.05–2.44) 0.87 (0.43–1.76) 0.73 (0.32–1.65) 
P for trend   0.207 0.127  0.595 0.060 
n > 0 501 20 0.78 (0.47–1.27) 0.65 (0.36–1.18) 68 0.93 (0.71–1.23) 0.72 (0.52–0.99) 
OSPFS
Fever frequency (n times/yr)TotalEventsHR (95% CI)aHR (95% CI)bEventsHR (95% CI)aHR (95% CI)b
Prediagnosis 
n = 0 2,240 143 1.00 (reference) 1.00 (reference) 298 1.00 (reference) 1.00 (reference) 
 0 < n ≤ 2 times/yr 1,896 148 0.89 (0.70–1.12) 0.82 (0.61–1.10) 286 0.87 (0.74–1.03) 0.81 (0.66–0.99) 
n > 2 times/yr 87 0.73 (0.30–1.77) 0.59 (0.15–2.40) 12 0.88 (0.50–1.58) 0.77 (0.34–1.73) 
P for trend   0.249 0.143  0.115 0.041 
n > 0 1983 153 0.88 (0.70–1.11) 0.81 (0.61–1.09) 298 0.87 (0.74–1.03) 0.81 (0.66–0.99) 
Postdiagnosis 
n = 0 2,056 75 1.00 (reference) 1.00 (reference) 237 1.00 (reference) 1.00 (reference) 
 0 < n ≤ 2 times/yr 445 19 0.85 (0.51–1.41) 0.71 (0.38–1.30) 60 0.94 (0.71–1.26) 0.72 (0.51–1.01) 
n > 2 times/yr 56 0.30 (0.04–2.13) 0.34 (0.05–2.44) 0.87 (0.43–1.76) 0.73 (0.32–1.65) 
P for trend   0.207 0.127  0.595 0.060 
n > 0 501 20 0.78 (0.47–1.27) 0.65 (0.36–1.18) 68 0.93 (0.71–1.23) 0.72 (0.52–0.99) 

aThe univariate COX model.

bAdjusted for age at diagnosis, education, menopausal status, clinical stage, ER status, HER2 status, and Ki-67.

Association of postdiagnostic infection-induced fever with breast cancer survival

The proportion of postdiagnostic infection-induced fever was 19.6% (501 patients). Over a median of 52.54 months of follow-up, 305 patients experienced breast cancer progression. As shown in Table 2, patients with postdiagnostic infection-induced fever had a significantly decreased risk of progression (HR 0.72; 95% CI, 0.52–0.99). However, there was a nonsignificant association between postdiagnostic infection-induced fever and OS (HR 0.65; 95% CI, 0.36–1.18; Table 2).

Interaction between IL6 rs1800796 and pre-/postdiagnostic infection-induced fever on the prognosis of breast cancer

The nonsignificant associations of IL6 rs1800796 with breast cancer OS and PFS were shown in Table 3. However, after adjusting with potential prognostic factors of breast cancer, the effect of prediagnostic infection-induced fever on breast cancer OS appeared to be opposite between CC genotype (HR 0.59; 95% CI, 0.36–0.99) and CG/GG genotype (HR 1.61; 95% CI, 0.83–3.15), accompanying with significant interaction (P = 0.021). Prediagnostic infection-induced fever also significantly decreased the risk of progression of breast cancer among women with the CC genotype of IL6 rs1800796 (HR 0.53; 95% CI, 0.36–0.79) but not among those with CG/GG genotypes (HR 0.85; 95% CI, 0.52–1.37). The interactions between postdiagnostic infection-induced fever and IL6 rs1800796 on survival were not significant (P > 0.05).

Table 3.

Association between pre/postdiagnostic infection-induced fever and prognosis of breast cancer stratified by IL6 rs1800796.

Overall survivalPFS
IL6 rs1800796FeverTotalaDeath (%)HR (95% CI)bHR (95% CI)cProgression (%)HR (95% CI)bHR (95% CI)c
All patients 
 CC — 697 81 (11.6) 1.00 (reference) 1.00 (reference) 127 (18.2) 1.00 (reference) 1.00 (reference) 
 CG/GG — 476 66 (13.9) 1.20 (0.87–1.67) 1.16 (0.80–1.68) 104 (21.8) 1.21 (0.93–1.56) 1.17 (0.88–1.56) 
Prediagnostic fever 
 CC No 230 33 (14.3) 1.00 (reference) 1.00 (reference) 55 (23.9) 1.00 (reference) 1.00 (reference) 
 Yes 467 48 (10.3) 0.64 (0.41–1.00) 0.59 (0.36–0.99) 72 (15.4) 0.54 (0.38–0.77) 0.53 (0.36–0.79) 
 CG + GG No 139 15 (10.8) 1.00 (reference) 1.00 (reference) 30 (21.6) 1.00 (reference) 1.00 (reference) 
 Yes 337 51 (15.1) 1.28 (0.72–2.28) 1.61 (0.83–3.15) 74 (22.0) 0.88 (0.58–1.35) 0.85 (0.52–1.37) 
P for interaction    0.045 0.021  0.057 0.130 
Postdiagnostic fever 
 CC No 338 20 (5.9) 1.00 (reference) 1.00 (reference) 45 (13.3) 1.00 (reference) 1.00 (reference) 
 Yes 142 4 (2.8) 0.45 (0.16–1.33) 0.50 (0.17–1.50) 19 (13.4) 0.98 (0.57–1.68) 0.99 (0.57–1.72) 
 CG + GG No 242 18 (7.4) 1.00 (reference) 1.00 (reference) 44 (18.2) 1.00 (reference) 1.00 (reference) 
 Yes 102 6 (5.9) 0.78 (0.31–1.96) 0.67 (0.23–1.92) 14 (13.7) 0.72 (0.40–1.32) 0.47 (0.23–0.98) 
P for interaction    0.448 0.592  0.459 0.143 
Overall survivalPFS
IL6 rs1800796FeverTotalaDeath (%)HR (95% CI)bHR (95% CI)cProgression (%)HR (95% CI)bHR (95% CI)c
All patients 
 CC — 697 81 (11.6) 1.00 (reference) 1.00 (reference) 127 (18.2) 1.00 (reference) 1.00 (reference) 
 CG/GG — 476 66 (13.9) 1.20 (0.87–1.67) 1.16 (0.80–1.68) 104 (21.8) 1.21 (0.93–1.56) 1.17 (0.88–1.56) 
Prediagnostic fever 
 CC No 230 33 (14.3) 1.00 (reference) 1.00 (reference) 55 (23.9) 1.00 (reference) 1.00 (reference) 
 Yes 467 48 (10.3) 0.64 (0.41–1.00) 0.59 (0.36–0.99) 72 (15.4) 0.54 (0.38–0.77) 0.53 (0.36–0.79) 
 CG + GG No 139 15 (10.8) 1.00 (reference) 1.00 (reference) 30 (21.6) 1.00 (reference) 1.00 (reference) 
 Yes 337 51 (15.1) 1.28 (0.72–2.28) 1.61 (0.83–3.15) 74 (22.0) 0.88 (0.58–1.35) 0.85 (0.52–1.37) 
P for interaction    0.045 0.021  0.057 0.130 
Postdiagnostic fever 
 CC No 338 20 (5.9) 1.00 (reference) 1.00 (reference) 45 (13.3) 1.00 (reference) 1.00 (reference) 
 Yes 142 4 (2.8) 0.45 (0.16–1.33) 0.50 (0.17–1.50) 19 (13.4) 0.98 (0.57–1.68) 0.99 (0.57–1.72) 
 CG + GG No 242 18 (7.4) 1.00 (reference) 1.00 (reference) 44 (18.2) 1.00 (reference) 1.00 (reference) 
 Yes 102 6 (5.9) 0.78 (0.31–1.96) 0.67 (0.23–1.92) 14 (13.7) 0.72 (0.40–1.32) 0.47 (0.23–0.98) 
P for interaction    0.448 0.592  0.459 0.143 

aThe number may not equal to the total number due to missing data.

bThe univariate COX model.

cAdjusted for age at diagnosis, menopausal status, ER status, HER2 status, tumor size, and nodal status.

Interaction between nodal status and pre-/postdiagnostic infection-induced fever on the prognosis of breast cancer

Subsequently, we attempted to examine whether the clinicopathologic characteristics affect the association of pre-/postdiagnostic infection-induced fever with breast cancer prognosis. Similar associations between pre-/postdiagnostic infection-induced fever and increased survival in breast cancer were observed across menopausal status, ER status, Ki-67, tumor size, and clinical stage (Supplementary Tables S1 and S2). Notably, postdiagnostic infection-induced fever significantly improved the PFS (HR 0.51; 95% CI, 0.32–0.81) among lymph node–positive breast cancer patients but not among lymph node–negative breast cancer patients (HR 1.12; 95% CI, 0.70–1.79; Table 4); the interaction was statistically significant (P = 0.020). Contrarily, this interaction was not observed for prediagnostic infection-induced fever. Similar effects of postdiagnostic infection-induced fever on OS were observed for lymph node–positive and lymph node–negative women (HR 0.45; 95% CI, 0.21–0.97; HR 1.27; 95% CI, 0.48–3.34, respectively; Table 4).

Table 4.

Association between pre/postdiagnostic infection-induced fever and prognosis of breast cancer stratified by nodal status.

OSPFS
Nodal statusFeverTotalDeath (%)HR (95% CI) aProgression (%)HR (95% CI) a
Prediagnosis 
 Negative No 1,166 37 (3.2) 1.00 (reference) 91 (7.8) 1.00 (reference) 
 Yes 1,040 40 (3.8) 0.74 (0.43–1.27) 100 (9.6) 0.82 (0.59–1.15) 
 Positive No 937 88 (9.4) 1.00 (reference) 175 (18.7) 1.00 (reference) 
 Yes 827 93 (11.2) 0.87 (0.61–1.24) 168 (20.3) 0.82 (0.64–1.06) 
P for interaction    0.560  0.966 
Postdiagnosis 
 Negative No 1,131 23 (2.0) 1.00 (reference) 87 (7.7) 1.00 (reference) 
 Yes 258 7 (2.7) 1.27 (0.48–3.34) 29 (11.2) 1.12 (0.70–1.79) 
 Positive No 798 47 (5.9) 1.00 (reference) 135 (16.9) 1.00 (reference) 
 Yes 209 10 (4.8) 0.45 (0.21–0.97) 30 (14.4) 0.51 (0.32–0.81) 
P for interaction    0.205  0.020 
OSPFS
Nodal statusFeverTotalDeath (%)HR (95% CI) aProgression (%)HR (95% CI) a
Prediagnosis 
 Negative No 1,166 37 (3.2) 1.00 (reference) 91 (7.8) 1.00 (reference) 
 Yes 1,040 40 (3.8) 0.74 (0.43–1.27) 100 (9.6) 0.82 (0.59–1.15) 
 Positive No 937 88 (9.4) 1.00 (reference) 175 (18.7) 1.00 (reference) 
 Yes 827 93 (11.2) 0.87 (0.61–1.24) 168 (20.3) 0.82 (0.64–1.06) 
P for interaction    0.560  0.966 
Postdiagnosis 
 Negative No 1,131 23 (2.0) 1.00 (reference) 87 (7.7) 1.00 (reference) 
 Yes 258 7 (2.7) 1.27 (0.48–3.34) 29 (11.2) 1.12 (0.70–1.79) 
 Positive No 798 47 (5.9) 1.00 (reference) 135 (16.9) 1.00 (reference) 
 Yes 209 10 (4.8) 0.45 (0.21–0.97) 30 (14.4) 0.51 (0.32–0.81) 
P for interaction    0.205  0.020 

aAdjusted for age at diagnosis, education, menopausal status, clinical stage, ER status, HER2 status, and Ki-67.

Association between changing patterns of pre-/postdiagnostic infection-induced fever and breast cancer prognosis

In this study, 45.3% (1,158/2,557) of breast cancer patients altered their fever status following their diagnosis (Supplementary Table S3). Compared with those who never experienced infection-induced fever before and after diagnosis, patients with only prediagnostic infection-induced fever had better survival [OS (HR 0.46; 95% CI, 0.26–0.81); PFS (HR 0.64; 95% CI, 0.48–0.87)], while the patients with only postdiagnostic infection-induced fever were associated with a nonsignificantly decreased risk of progression [OS (HR 0.98; 95% CI, 0.50–1.95); PFS (HR 0.74; 95% CI, 0.47–1.18)]. Notably, the prognosis was improved significantly when the patients experienced infection-induced fever before and after diagnosis [OS (HR 0.15; 95% CI, 0.05–0.49); PFS (HR 0.48; 95% CI, 0.30–0.75); Table 5].

Table 5.

The relationship between febrile status and prognosis before and after diagnosis of breast cancer.

Overall survivalPFS
Prediagnostic feverPostdiagnostic feverTotalEventsHR (95% CI)aEventsHR (95% CI)a
No No 1,117 44 1.00 (reference) 134 1.00 (reference) 
Yes No 939 31 0.46 (0.26–0.81) 103 0.64 (0.48–0.87) 
No Yes 219 14 0.98 (0.50–1.95) 31 0.74 (0.47–1.18) 
Yes Yes 282 0.15 (0.05–0.49) 37 0.48 (0.30–0.75) 
Overall survivalPFS
Prediagnostic feverPostdiagnostic feverTotalEventsHR (95% CI)aEventsHR (95% CI)a
No No 1,117 44 1.00 (reference) 134 1.00 (reference) 
Yes No 939 31 0.46 (0.26–0.81) 103 0.64 (0.48–0.87) 
No Yes 219 14 0.98 (0.50–1.95) 31 0.74 (0.47–1.18) 
Yes Yes 282 0.15 (0.05–0.49) 37 0.48 (0.30–0.75) 

aAdjusted for age at diagnosis, education, menopausal status, clinical stage, ER status, HER2 status, and ki-67.

We conducted a large prospective cohort study to examine a potential protective effect of infection-induced fever on breast cancer survival. Our findings suggested a possible association between infection-induced fever and increased breast cancer survival. Quite a lot of previous studies supported this result (29). For example, Ruckdeschel and colleagues found that the 5-year survival rate for patients who developed empyema after lung cancer was 50% as compared with 18% for noninfected patients (30); Jeys and colleagues found evidence for increased survival in osteosarcoma patients with postoperative infection (31); Pasquale and colleagues reported that glioblastoma patients without a postoperative infection showed an adjusted HR for overall survival of 2.3 (95% CI, 1.0–5.3; ref. 32). Furthermore, we firstly found the protective effect on breast cancer prognosis in combination with pre- and postdiagnostic infection-induced fever (previous studies only evaluated the effect of postdiagnostic fever).

The mechanisms underlying fever improving the prognosis of breast cancer were not clear and might be related to the stimulation of innate and adaptive immune responses during infection (1). Fever would be able to greatly activate adaptive immune responses, such as stimulating CD8+ effector T cell, augmenting the cytolytic activity of NK cells, and enhancing the phagocytic potential of macrophages and dendritic cells (DC; refs. 33–35). All of these responses could significantly suppress the tumor progression. In addition, it was observed that breast cancer cells are more fragile and vulnerable to heat with apoptosis than normal cells (36). The necrotic or heat-stressed cancer cells would supply more tumor antigens and activate the antitumor immune responses (37).

Consistent with our study, several previous population studies did not find a significant association between the IL6 rs1800796 and survival (24, 25). However, we found that the decreased risk of progression by prediagnostic infection-induced fever was more evident among women with the CC genotype than those with CG/GG genotypes of IL6 rs1800796. Several studies reported that the rs1800796 C allele was associated with increased levels of IL6 (38, 39). Additionally, the IL6 levels were heritable with estimates from twin studies from 15% to 61% (40–43), which indicated that the IL6 SNPs might be a reliable surrogate marker for IL6 levels. After stratifying by IL6 rs1800796, breast cancer patients could be divided into two subgroups: high level of the IL6 (i.e., CC genotype) group and low level of the IL6 (i.e., CG/GG genotype) group. IL6 was an important mediator between fever and antitumor immune response (1). For example, after stimulating by fever, the breast cancer patients could boost antitumor T-cell trafficking into lymph nodes through IL6 trans-signaling, and then the proliferation of the breast cancer cells would be suppressed (15). In addition, fever could also increase the levels of IL6 (44), which might activate the signal transducer and activator of transcription 3 (STAT3) in the downstream signaling (45). It has been found that the nuclear expression of phosphorylated STAT3 (pSTAT3, the activated form of STAT3) was associated with small tumor size, low grade, negative lymphovascular invasion, and improved breast cancer–specific survival (46). Thus, among the patients with CC genotype, infection-induced fever may be easier to exert antitumor effects via sufficient IL6 and STAT3 activation, resulting in that febrile women with CC genotype had a significantly decreased risk of progression of breast cancer. On the contrary, the infection-induced fever would not be able to boost the effective immune response to suppress tumor progression in carriers with the G allele because of insufficient IL6, indicating that the contribution to survival by infection-induced fever may be varied by the different genotype of IL6 rs1800796.

Interestingly, we further found that the protective effects of postdiagnostic infection-induced fever were more evident among patients with lymph node–positive than those with lymph node–negative. Lu and colleagues also found that postoperative fever was associated with an increased rate of progression in node-negative breast cancer patients (13). Fever could augment the expression of IL6 (20, 21), which might boost T cells with antitumor properties trafficking to lymph nodes and tumor sites (18). Compared with tumor-free nodes, tumor-involved nodes might accumulate more antitumor T cells. Thus, it would be reasonable that postdiagnostic febrile patients had a better prognosis compared with nonfebrile ones among the lymph node–positive patients. For lymph node–negative patients, there were no tumor cells in the lymph nodes, so the antitumor immune responses induced by infection-induced fever in febrile patients were not different from nonfebrile patients.

Nevertheless, several limitations of this study should be mentioned. Firstly, our measures of infection-induced fever were self-reported, which inevitably caused a recall bias. However, the patients in this study were unaware of the hypothesis, and misclassification due to recall bias might occur equally in patients with good and poor prognoses. This nondifferential exposure misclassification was likely to get biased toward the null, and the effects were possibly underestimated (47). Secondly, selection bias might exist because 298 patients with missing prediagnostic infection-induced fever records were excluded. However, clinicopathologic characteristics were equally distributed between patients with infection-induced fever records and without infection-induced fever records (Supplementary Table S4). Thirdly, the present study did not evaluate IL6 levels. However, IL6 levels vary greatly depending on the time of sample collection, sample type, food intake, and stress (48). Stable genomic information on cytokines might be better for assessing ongoing systemic inflammation because it was not subject to daily fluctuations. Fourthly, we evaluated the modification effect of IL6 rs1800796 among only a subset of 1,173 individuals, and the results may not be true for the whole study population. However, we found that the main effects in patients with IL6 SNP data were the same as those in all patients (Supplementary Table S5); moreover, the patients with and without IL6 SNP data were from the same area of Guangzhou, and the genotype distribution should not be different between the two subgroups, indicating that the modification effect would maintain in the whole study population. Nevertheless, validation studies in other populations are needed. Fifthly, there was possibly collider stratification bias, i.e., prediagnostic infection-induced fever → breast cancer risk ← risk/prognostic factors. In the absence of appropriate controls (noncases) in this study, we could not mitigate this potential bias using the inverse probability of censoring weights (49). However, this bias might only impact the generalizability to other populations (50). Finally, missing information on treatment and socioeconomic status in our study might confound the results; however, the adjustment for clinicopathologic characteristics and education level partly compensated for this defect.

In conclusion, it was found that prediagnostic and postdiagnostic infection-induced fever improved the prognosis of breast cancer patients, particularly for CC genotype of IL6 rs1800796 or lymph node–positive patients. These findings provide an insight into the role of infection-induced fever as a marker of breast cancer prognosis and indicate that fever under medical guidance may be considered as a part of the therapy regimen for breast cancer. Further studies are warranted to explore the underlying mechanisms.

H. Ye reports grants from the Science and Technology Planning Project of Guangdong Province, China, the National Natural Science Foundation of China, and the National Natural Science Foundation of China during the conduct of the study. Z.-Z. Liang reports grants from the Science and Technology Planning Project of Guangdong Province, China, the National Natural Science Foundation of China, and from the National Natural Science Foundation of China during the conduct of the study. Y. Li reports grants from the Science and Technology Planning Project of Guangdong Province, China, the National Natural Science Foundation of China, and the National Natural Science Foundation of China during the conduct of the study. Q. Liu reports grants from the Science and Technology Planning Project of Guangdong Province, China, the National Natural Science Foundation of China, and the National Natural Science Foundation of China during the conduct of the study. Y. Lin reports grants from the Science and Technology Planning Project of Guangdong Province, China, the National Natural Science Foundation of China, and the National Natural Science Foundation of China during the conduct of the study. Z. Ren reports grants from the Science and Technology Planning Project of Guangdong Province, China, the National Natural Science Foundation of China, and the National Natural Science Foundation of China during the conduct of the study. No disclosures were reported by the other authors.

H. Ye: Conceptualization, data curation, formal analysis, investigation, methodology, writing–original draft, writing–review and editing. L.-Y. Tang: Formal analysis. Z.-Z. Liang: Formal analysis. Q.-X. Chen: Formal analysis. Y.-Q. Li: Data curation. Q. Liu: Resources, data curation. X. Xie: Resources, data curation. Y. Lin: Resources, supervision, project administration. Z.-F. Ren: Conceptualization, resources, data curation, formal analysis, supervision, writing–original draft, project administration, writing–review and editing.

We sincerely thank the patients who participated in this study, the staff who conducted the baseline and the follow‐up data collection, and the medical staff in the breast departments of the First Affiliated Hospital, the Second Affiliated Hospital, and the Cancer Center of Sun Yat‐sen University. This work was supported by the Science and Technology Planning Project of Guangdong Province, China (grant 2019B030316002) and the National Natural Science Foundation of China (grants 81773515 and 81973115).

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Note: Supplementary data for this article are available at Cancer Epidemiology, Biomarkers & Prevention Online (http://cebp.aacrjournals.org/).

1.
Evans
SS
,
Repasky
EA
,
Fisher
DT
.
Fever and the thermal regulation of immunity: the immune system feels the heat
.
Nat Rev Immunol
2015
;
15
:
335
49
.
2.
Hoption Cann
SA
,
van Netten
JP
,
van Netten
C
.
Acute infections as a means of cancer prevention: opposing effects to chronic infections?
Cancer Detect Prev
2006
;
30
:
83
93
.
3.
Kucerova
P
,
Cervinkova
M
.
Spontaneous regression of tumour and the role of microbial infection–possibilities for cancer treatment
.
Anticancer Drugs
2016
;
27
:
269
77
.
4.
Hoption Cann
SA
,
van Netten
JP
,
van Netten
C
,
Glover
DW
.
Spontaneous regression: a hidden treasure buried in time
.
Med Hypotheses
2002
;
58
:
115
9
.
5.
McCarthy
EF
.
The toxins of William B. Coley and the treatment of bone and soft-tissue sarcomas
.
Iowa Orthop J
2006
;
26
:
154
8
.
6.
Challenor
S
,
Tucker
D
.
SARS-CoV-2-induced remission of Hodgkin lymphoma
.
Br J Haematol
2021
;
192
:
415
.
7.
Qin
L
,
Chen
C
,
Chen
L
,
Xue
R
,
Ou-Yang
M
,
Zhou
C
, et al
.
Worldwide malaria incidence and cancer mortality are inversely associated
.
Infectious agents and cancer
2017
;
12
:
14
.
8.
Sung
H
,
Ferlay
J
,
Siegel
RL
,
Laversanne
M
,
Soerjomataram
I
,
Jemal
A
, et al
.
Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries
.
CA Cancer J Clin
2021
;
71
:
209
49
.
9.
Łukasiewicz
S
,
Czeczelewski
M
,
Forma
A
,
Baj
J
,
Sitarz
R
,
Stanisławek
A
.
Breast cancer-epidemiology, risk factors, classification, prognostic markers, and current treatment strategies-an updated review
.
Cancers
2021
;
13
:
4287
.
10.
Morra
A
,
Jung
AY
,
Behrens
S
,
Keeman
R
,
Ahearn
TU
,
Anton-Culver
H
, et al
.
Breast cancer risk factors and survival by tumor subtype: pooled analyses from the breast cancer association consortium
.
Cancer Epidemiol Biomarkers Prev
2021
;
30
:
623
42
.
11.
Su
Y
,
Tang
LY
,
Chen
LJ
,
He
JR
,
Su
FX
,
Lin
Y
, et al
.
Joint effects of febrile acute infection and an interferon-γ polymorphism on breast cancer risk
.
PLoS One
2012
;
7
:
e37275
.
12.
Teucher
G
,
Schindler
AE
.
[Postoperative fever and prognosis in breast cancer]
.
Arch Geschwulstforsch
1987
;
57
:
309
17
.
13.
Yan
T
,
Yin
W
,
Zhou
L
,
Jiang
Y
,
Shen
Z
,
Shao
Z
, et al
.
Postoperative fever: the potential relationship with prognosis in node negative breast cancer patients
.
PLoS One
2010
;
5
:
e15903
.
14.
Shen
RN
,
Hornback
NB
,
Shidnia
H
,
Shupe
RE
,
Brahmi
Z
.
Whole-body hyperthermia decreases lung metastases in lung tumor-bearing mice, possibly via a mechanism involving natural killer cells
.
J Clin Immunol
1987
;
7
:
246
53
.
15.
Fisher
DT
,
Chen
Q
,
Skitzki
JJ
,
Muhitch
JB
,
Zhou
L
,
Appenheimer
MM
, et al
.
IL-6 trans-signaling licenses mouse and human tumor microvascular gateways for trafficking of cytotoxic T cells
.
J Clin Invest
2011
;
121
:
3846
59
.
16.
Chen
Q
,
Fisher
DT
,
Clancy
KA
,
Gauguet
JM
,
Wang
WC
,
Unger
E
, et al
.
Fever-range thermal stress promotes lymphocyte trafficking across high endothelial venules via an interleukin 6 trans-signaling mechanism
.
Nat Immunol
2006
;
7
:
1299
308
.
17.
Kang
S
,
Tanaka
T
,
Narazaki
M
,
Kishimoto
T
.
Targeting interleukin-6 signaling in clinic
.
Immunity
2019
;
50
:
1007
23
.
18.
Fisher
DT
,
Appenheimer
MM
,
Evans
SS
.
The two faces of IL-6 in the tumor microenvironment
.
Semin Immunol
2014
;
26
:
38
47
.
19.
Ostberg
JR
,
Taylor
SL
,
Baumann
H
,
Repasky
EA
.
Regulatory effects of fever-range whole-body hyperthermia on the LPS-induced acute inflammatory response
.
J Leukocyte Biol
2000
;
68
:
815
20
.
20.
Jiang
Q
,
Detolla
L
,
Singh
IS
,
Gatdula
L
,
Fitzgerald
B
,
van Rooijen
N
, et al
.
Exposure to febrile temperature upregulates expression of pyrogenic cytokines in endotoxin-challenged mice
.
Am J Physiol
1999
;
276
:
R1653
60
.
21.
Duan
XH
,
Li
H
,
Han
XW
,
Ren
JZ
,
Li
FY
,
Ju
SG
, et al
.
Upregulation of IL-6 is involved in moderate hyperthermia induced proliferation and invasion of hepatocellular carcinoma cells
.
Eur J Pharmacol
2018
;
833
:
230
6
.
22.
Hasday
JD
,
Thompson
C
,
Singh
IS
.
Fever, immunity, and molecular adaptations
.
Comprehensive Physiology
2014
;
4
:
109
48
.
23.
Cartmell
T
,
Poole
S
,
Turnbull
AV
,
Rothwell
NJ
,
Luheshi
GN
.
Circulating interleukin-6 mediates the febrile response to localised inflammation in rats
.
J Physiol
2000
;
526
Pt 3(Pt 3)
:
653
61
.
24.
DeMichele
A
,
Gray
R
,
Horn
M
,
Chen
J
,
Aplenc
R
,
Vaughan
WP
, et al
.
Host genetic variants in the interleukin-6 promoter predict poor outcome in patients with estrogen receptor-positive, node-positive breast cancer
.
Cancer Res
2009
;
69
:
4184
91
.
25.
Markkula
A
,
Simonsson
M
,
Ingvar
C
,
Rose
C
,
Jernström
H
.
IL6 genotype, tumour ER-status, and treatment predicted disease-free survival in a prospective breast cancer cohort
.
BMC Cancer
2014
;
14
:
759
.
26.
He
JR
,
Tang
LY
,
Yu
DD
,
Su
FX
,
Song
EW
,
Lin
Y
, et al
.
Epstein-Barr virus and breast cancer: serological study in a high-incidence area of nasopharyngeal carcinoma
.
Cancer Lett
2011
;
309
:
128
36
.
27.
Liao
YH
,
Ren
JT
,
Zhang
W
,
Zhang
ZZ
,
Lin
Y
,
Su
FX
, et al
.
Polymorphisms in homologous recombination repair genes and the risk and survival of breast cancer
.
J Gene Med
2017
;
19
:
2988
.
28.
Zhu
RM
,
Lin
W
,
Zhang
W
,
Ren
JT
,
Su
Y
,
He
JR
, et al
.
Modification effects of genetic polymorphisms in FTO, IL-6, and HSPD1 on the associations of diabetes with breast cancer risk and survival
.
PLoS One
2017
;
12
:
e0178850
.
29.
Hobohm
U
.
Fever and cancer in perspective
.
Cancer Immunol Immunother
2001
;
50
:
391
6
.
30.
Ruckdeschel
JC
,
Codish
SD
,
Stranahan
A
,
McKneally
MF
.
Postoperative empyema improves survival in lung cancer. Documentation and analysis of a natural experiment
.
N Engl J Med
1972
;
287
:
1013
7
.
31.
Jeys
LM
,
Grimer
RJ
,
Carter
SR
,
Tillman
RM
,
Abudu
A
.
Postoperative infection and increased survival in osteosarcoma patients: are they associated?
Ann Surg Oncol
2007
;
14
:
2887
95
.
32.
De Bonis
P
,
Albanese
A
,
Lofrese
G
,
de Waure
C
,
Mangiola
A
,
Pettorini
BL
, et al
.
Postoperative infection may influence survival in patients with glioblastoma: simply a myth?
Neurosurgery
2011
;
69
:
864
8
;
discussion 8–9
.
33.
O'Sullivan
D
,
Stanczak
MA
,
Villa
M
,
Uhl
FM
,
Corrado
M
,
Klein Geltink
RI
, et al
.
Fever supports CD8(+) effector T cell responses by promoting mitochondrial translation
.
Proc Nat Acad Sci USA
2021
;
118
.
34.
Ostberg
JR
,
Dayanc
BE
,
Yuan
M
,
Oflazoglu
E
,
Repasky
EA
.
Enhancement of natural killer (NK) cell cytotoxicity by fever-range thermal stress is dependent on NKG2D function and is associated with plasma membrane NKG2D clustering and increased expression of MICA on target cells
.
J Leukocyte Biol
2007
;
82
:
1322
31
.
35.
Bachleitner-Hofmann
T
,
Strohschneider
M
,
Krieger
P
,
Sachet
M
,
Dubsky
P
,
Hayden
H
, et al
.
Heat shock treatment of tumor lysate-pulsed dendritic cells enhances their capacity to elicit antitumor T cell responses against medullary thyroid carcinoma
.
J Clin Endocrinol Metab
2006
;
91
:
4571
7
.
36.
Amaya
C
,
Kurisetty
V
,
Stiles
J
,
Nyakeriga
AM
,
Arumugam
A
,
Lakshmanaswamy
R
, et al
.
A genomics approach to identify susceptibilities of breast cancer cells to “fever-range” hyperthermia
.
BMC Cancer
2014
;
14
:
81
.
37.
Hobohm
U
,
Stanford
JL
,
Grange
JM
.
Pathogen-associated molecular pattern in cancer immunotherapy
.
Crit Rev Immunol
2008
;
28
:
95
107
.
38.
Gu
W
,
Du
DY
,
Huang
J
,
Zhang
LY
,
Liu
Q
,
Zhu
PF
, et al
.
Identification of interleukin-6 promoter polymorphisms in the Chinese Han population and their functional significance
.
Crit Care Med
2008
;
36
:
1437
43
.
39.
Brull
DJ
,
Montgomery
HE
,
Sanders
J
,
Dhamrait
S
,
Luong
L
,
Rumley
A
, et al
.
Interleukin-6 gene -174g>c and -572g>c promoter polymorphisms are strong predictors of plasma interleukin-6 levels after coronary artery bypass surgery
.
Arterioscler Thromb Vasc Biol
2001
;
21
:
1458
63
.
40.
Wörns
MA
,
Victor
A
,
Galle
PR
,
Höhler
T
.
Genetic and environmental contributions to plasma C-reactive protein and interleukin-6 levels–a study in twins
.
Genes Immun
2006
;
7
:
600
5
.
41.
Sas
AA
,
Jamshidi
Y
,
Zheng
D
,
Wu
T
,
Korf
J
,
Alizadeh
BZ
, et al
.
The age-dependency of genetic and environmental influences on serum cytokine levels: a twin study
.
Cytokine
2012
;
60
:
108
13
.
42.
Neijts
M
,
van Dongen
J
,
Kluft
C
,
Boomsma
DI
,
Willemsen
G
,
de Geus
EJ
.
Genetic architecture of the pro-inflammatory state in an extended twin-family design
.
Twin Res Hum Genet
2013
;
16
:
931
40
.
43.
Amaral
WZ
,
Krueger
RF
,
Ryff
CD
,
Coe
CL
.
Genetic and environmental determinants of population variation in interleukin-6, its soluble receptor and C-reactive protein: insights from identical and fraternal twins
.
Brain Behav Immun
2015
;
49
:
171
81
.
44.
Rao
L
,
Zhang
K
,
Luo
H
,
He
S
,
Li
Y
,
Liu
C
, et al
.
The ability of inflammatory markers to recognize infection in cancer patients with fever at admission
.
Immunol Res
2022 Jun 28 [Epub ahead of print]
.
45.
Braun
DA
,
Fribourg
M
,
Sealfon
SC
.
Cytokine response is determined by duration of receptor and signal transducers and activators of transcription 3 (STAT3) activation
.
J Biol Chem
2013
;
288
:
2986
93
.
46.
Aleskandarany
MA
,
Agarwal
D
,
Negm
OH
,
Ball
G
,
Elmouna
A
,
Ashankyty
I
, et al
.
The prognostic significance of STAT3 in invasive breast cancer: analysis of protein and mRNA expressions in large cohorts
.
Breast Cancer Res Treat
2016
;
156
:
9
20
.
47.
Jurek
AM
,
Greenland
S
,
Maldonado
G
,
Church
TR
.
Proper interpretation of non-differential misclassification effects: expectations vs observations
.
Int J Epidemiol
2005
;
34
:
680
7
.
48.
Zhou
X
,
Fragala
MS
,
McElhaney
JE
,
Kuchel
GA
.
Conceptual and methodological issues relevant to cytokine and inflammatory marker measurements in clinical research
.
Curr Opin Clin Nutr Metab Care
2010
;
13
:
541
7
.
49.
Robins
JM
,
Finkelstein
DM
.
Correcting for noncompliance and dependent censoring in an AIDS Clinical Trial with inverse probability of censoring weighted (IPCW) log-rank tests
.
Biometrics
2000
;
56
:
779
88
.
50.
Lash
TL
,
Rothman
KJ
.
Selection bias and generalizability
.
In
:
Lash
TL
,
VanderWeele
TJ
,
Haneuse
S
,
Rothman
KJ
,
editors
.
Modern epidemiology
, 4th edition. Philadelphia, PA:
Wolters Kluwer
;
2021
.

Supplementary data