You are here

The genetic variants underlying breast cancer treatment-induced chronic and late toxicities: A systematic review

Cancer Treatment Reviews, 10, 40, pages 1199 - 1214



  • Breast cancer treatment may cause several undesirable chronic and late toxicities.
  • Toxicities include cardiotoxicity, cognitive dysfunction, fatigue and neuropathy.
  • There may be a genetic component underlying these treatment-induced toxicities.
  • Certain genetic variants are associated with the four toxicities.
  • Methodological concerns in the studies may threaten the validity of the findings.


A systematic review was performed to describe the findings from 19 genetic association studies that have examined the genetic variants underlying four common treatment-induced chronic and late toxicities in breast cancer patients, and to evaluate the quality of reporting. Three out of 5 studies found an association betweenHER2lle655Val polymorphisms and trastuzumab-induced cardiotoxicity. Two studies found a positive association between cognitive impairment and the Val allele of theCOMTgene and the ε4 allele of the apolipoprotein E gene. Genetic associations were established between fatigue and the G/G genotype ofIL6-174andTNF-308, and the Met allele of theCOMTgene in 4 studies. Among studies (N = 8) that evaluated the genetic associations underlying peripheral neuropathy, CYP2C8∗3 variant is commonly reported as the associated gene. Most studies failed to conform to the major criteria listed in the STREGA guidelines, with a lack of transparent reporting of methods and results.

Keywords: Breast cancer, Cardiotoxicity, Cognitive dysfunction, Fatigue, Peripheral neuropathy, Genetic association.


With the advent of therapeutics and the improved early detection of malignancies, the number of breast cancer survivors is predicted to increase over the next decade [1] . Toxicities from anti-cancer treatments are detrimental and patients with breast cancer are often plagued by a wide range of treatment-induced toxicities that lead to functional impairment with great economic, emotional and social cost[2] and [3]. While toxicities such as nausea and vomiting occur acutely during treatment, toxicities that occur during treatment and persist beyond cancer recovery are known as chronic toxicities, whereas toxicities that have a delayed onset, typically months or years post-cancer treatment, are known as late toxicities [1] .

Cardiotoxicity, cognitive dysfunction, fatigue and peripheral neuropathy are among the chronic and late toxicities induced by breast cancer treatments. Among them, cognitive dysfunction, fatigue and peripheral neuropathy are physical and treatment-related adverse effects that are commonly experienced by breast cancer patients and survivors[2] and [3]. Cardiotoxicity can manifest as mild transient electrocardiographic changes or serious arrhythmias and cardiomyopathy. Patients manifesting peripheral neuropathy may suffer from burning or tingling sensations in their extremities. Its occurrence can lead to an irreversible loss of function of the peripheral nerves[4], [5], and [6]. On the other hand, fatigue and cognitive dysfunction have detrimental effects on quality of life and daily functioning in survivors of breast cancer. Fatigue refers to a distressing, persistent, subjective sense of physical, emotional and/or cognitive tiredness. Cognitive dysfunction encompasses a wide range of subtle cognitive changes, such as memory loss, the inability to concentrate and difficulty in thinking[7], [8], [9], and [10]. Unfortunately, the mechanisms underlying these chronic and late toxicities remain poorly understood. Although clinical and demographic factors may predispose certain individuals to greater risk and severity of toxicities, the vast inter-individual variation in the manifestation of the toxicities accentuates the idiosyncratic nature of these toxicities. Genetics may therefore provide greater insights into the development of these cancer treatment-induced toxicities. By identifying the functional genetic variants associated with these toxicities, we are able to improve our understanding of the underlying mechanistic pathways, in order to path the way for development of novel therapies targeting these toxicities. Moreover, the identification of genetic biomarkers may identify patients at prior to treatment, which will allow the use of suitable therapeutic measures to prevent and mitigate these toxicities.

The literature provides important insights into the role of genetics in these toxicities. Recent observations on paclitaxel metabolism through the CYP2C8 enzyme system suggest that variations in theCYP2C8genotype may increase the risk of peripheral neuropathy via the attenuation of paclitaxel metabolism[11] and [12]. With a greater exposure to paclitaxel, severe forms of paclitaxel-induced peripheral neuropathy may be induced due to axon degeneration and lead to the persistence of the neurotoxicity[13], [14], [15], and [16]. A genome-wide association study provided evidence suggesting that genetic variations in the ephrin type A receptor (EPHA) genes are strongly associated with paclitaxel-induced peripheral neuropathy [17] . The EPHA receptors are involved in axonal guidance and regeneration following injury. It has been suggested thatEPHAgenetic variants may contribute to peripheral neuropathy by interfering with axon repair. The role of pro-inflammatory cytokine gene polymorphisms in toxicities such as fatigue and cognitive disturbances has been increasingly investigated. Studies have shown that single-nucleotide polymorphisms (SNPs) in the regulatory regions of the cytokine genes can affect the quantitative expression of cytokines[18] and [19]. This may result in higher circulating levels of cytokines and increase an individual’s susceptibility to these undesirable symptoms.

The role of genetic variants in the four toxicities has been increasingly examined and an understanding of its contribution is necessary for uncovering the biological relationships between functional genetic variants and treatment-induced toxicities[20] and [21]. The quality of the reporting in the existing genetic association studies has not been investigated using robust evaluation criteria, such as the Strengthening the Reporting of Genetic Association Studies (STREGA) reporting guidelines. As such, the aim of this systematic review was to report and evaluate the findings of existing genetic association studies that have examined the genetic variants underlying chronic and late toxicities, specifically cardiotoxicity, cognitive dysfunction, fatigue and peripheral neuropathy, in patients treated with different modalities for breast cancer. The reporting quality of these studies was also assessed using the STREGA guidelines.


Literature search strategy

A systematic literature search was performed using PubMed and Scopus in May 2014. The search was limited to English language articles. There were no screening limits in terms of the publication dates. Additional studies were gleaned from the bibliographies of the first set of identified studies and from studies identified for review. The literature search was conducted using Medical Subject Headings and combinations of relevant keywords, such as “breast cancer,” “chemotherapy,” “cardiotoxicity,” “cardiovascular disease,” “chemobrain,” “cognitive impairment,” “cognitive dysfunction,” “fatigue,” “gene,” “genetic,” “gene polymorphism,” “hormonal therapy,” “neuropathy,” “peripheral neuropathy” and “targeted therapy.”

The literature search was performed by two investigators independently and was cross-checked for consistency to ensure its reliability and reproducibility.

Study inclusion–exclusion criteria

Published studies were included in the review if they evaluated at least one of the breast cancer treatment-induced toxicities: cardiotoxicity, cognitive dysfunction, fatigue or peripheral neuropathy. The study must evaluate a patient population that comprised at least 50% breast cancer patients and/or survivors; and studied a patient population treated with cytotoxic chemotherapy, targeted therapy or hormonal therapy, or a combination of these treatment modalities.

Interim studies, laboratory studies, case series and reports, systematic reviews and meta-analyses were excluded. The genetic associations underlying endocrine treatment-induced toxicities such as tamoxifen-induced hot flashes and aromatase inhibitor-associated arthralgia were not evaluated in this study, as extensive reviews were recently performed on these toxicities[22] and [23].

Data extraction

Two investigators independently extracted the data using a pre-designed, piloted form. Each study was reviewed and the relevant information was extracted and compiled. The following information was extracted from each study: (1) Study objectives, (2) Study design, (3) Characteristics of study participants, (4) Study’s inclusion and exclusion criteria, (5) Type of treatments study participants received, (6) Total number of study participants and number of blood samples genotyped, (7) Type of outcome measures and (8) Covariates that were controlled in the final statistical model.

Quality of reporting using the STREGA guidelines

The quality of the reporting in each study was evaluated using the STREGA reporting guidelines[24] and [25]. STREGA identifies five main categories of information that should be included in the reporting of genetic association studies, based on common problems identified in existing studies: the reporting of possible genotyping errors, the methods used to address population stratification, the methods used for the inference of haplotypes or genotypes, whether the Hardy–Weinberg equilibrium was considered and whether the study was the first report of a genetic association, a replication effort or both. Two investigators independently assessed the quality of the reporting and differences in the assessment were resolved by discussion or by a third person if consensus was not achieved by the two investigators.

Due to the heterogeneity in the study design, interventions and outcome measurements among the various reviewed studies, meta-analysis was not performed in this review.


Five hundred and twelve relevant studies were identified, with 19 studies met the predetermined inclusion criteria ( Fig. 1 ). The included studies examined the genetic associations underlying one of the four treatment-induced toxicities, cardiotoxicity (N = 5), cognitive dysfunction (N = 2), fatigue (N = 4) and peripheral neuropathy (N = 8)[11], [12], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], and [42]. Table 1 summarizes the genetic variants that were positively associated with each toxicity.


Fig. 1 Inclusion and exclusion process flowchart.

Table 1 Genetic variants that are positively associated with each toxicity.

Toxicity type Gene Nucleotide change Amino acid change (if any) Reference SNP/allele designation Types of effect (PD/PK/Physiological) References
Cardiotoxicity - HER2

1963 A > G Ile655Val rs1136201 PD [26], [27], and [28]
Cognitive dysfunction - APOE

7930 T > C

8041 C > T

rs429358, rs7412 (ε4 allele) Physiological [31]

472 G > A Val158Met rs4680 Physiological [32]
Fatigue - COMT

472 G > A Val158Met rs4680 Physiological [36]
- IL6

−174 G > C None rs1800795 Physiological [34]

−308 G > A None rs1800629 Physiological [34]
Peripheral neuropathy - ABCG1

Intronic 189+2659 A > G None rs492338 Physiological [38]

2677 T > G/A Ser893Ala/Thr rs2032582, PK [42]
−129 T > C None rs3213619

4544 G > A Cys1515Tyr rs8187710 PK [42]
- CYP2C8

416 G > A Arg139Lys Lys399Arg rs11572080, rs105096081 (3 allele) PK [11], [12], and [38]
1196 A > G
792 C > G Ile264Met 1058930 (4 allele) PK [42]
- CYP1B1

4326 C > G Leu432Val rs1056836 (3 allele) PK [42]

Intronic 97627774 G > A None rs301927 Physiological [42]

Intronic 10199 C > A None rs7648104, rs7637888, rs6786638, rs6442150 Physiological [41]
Intronic 17610 T > C
Intronic 54936 G > C
Intronic 59463 T > C
- FZD3

Intronic 16657 A > G None rs7001034 Physiological [37]

Intronic 918-24415 G > A None rs6473187 Physiological [42]

Intronic -61-2168 C > T None rs3829306 PK [42]

−157 A > G None rs9501929 PD [42]

Abbreviations:PD, pharmacodynamic; PK, pharmacokinetic.

The quality of the reporting in the studies ( Table 2 )

Among the reviewed studies, one study fulfilled the five main criteria in the STREGA guidelines for the reporting of data generated from genetic association studies. The majority of the studies (N = 13) did not report the error rates and call rates associated with their genotyping methods. Seventeen (89%) of the studies did not state whether their genotyping was performed in batches or simultaneously. Eight (42%) of the studies failed to provide any information on whether population stratification was assessed and corrected in their analyses.

Table 2 The quality of reporting in each study (N = 19).

Studies a 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Genotyping methods and errors
Describe the laboratory methods: state the source and storage of DNA, the genotyping methods and the platforms fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1
Describe the laboratory methods: state the error rates and call rates fx2 fx2 fx2 fx2 fx2 fx2 fx2 fx2 fx2 fx2 fx2 fx1 fx1 fx1 fx1 fx2 fx2 fx1 fx1
State the laboratory/center where the genotyping was done fx2 fx2 fx2 fx2 fx2 fx1 fx1 fx2 fx2 fx2 fx1 fx1 fx1 fx2 fx1 fx2 fx2 fx1 fx1
Specify whether genotypes were assigned using all of the data from the study simultaneously or in smaller batches fx2 fx2 fx2 fx2 fx2 fx2 fx2 fx2 fx2 fx1 fx2 fx2 fx2 fx2 fx2 fx2 fx2 fx2 fx1
Report the numbers of individuals for whom genotyping was attempted and the numbers of individuals for whom genotyping was successful fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx2 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1
Modelling population stratification
Describe any methods used to assess or address population stratification fx2 fx1 fx2 fx2 fx1 fx2 fx2 fx1 fx1 fx1 fx2 fx1 fx1 fx1 fx1 fx2 fx2 fx1 fx1
Modelling haplotype variation
Describe any methods used for inferring genotypes or haplotypes NA NA NA NA NA NA NA NA NA NA NA fx1 NA fx1 NA fx1 NA fx1 fx1
Hardy–Weinberg equilibrium
State whether the Hardy–Weinberg equilibrium was considered fx1 fx1 fx1 fx2 fx1 fx2 fx2 fx2 fx1 fx1 fx2 fx1 fx1 fx1 fx1 fx2 fx1 fx1 fx1
State if the study is the first report of a genetic association, a replication effort or both fx1 fx1 fx1 fx2 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx1 fx2 fx1 fx1 fx1

a 1. Lemieux J et al. Anticancer Res, 2013; 2. Beauclair S et al. Ann Oncol, 2007; 3. Roca L et al. Breast Cancer Res Treat, 2013; 4. Kitagawa K et al. Ann Oncol, 2012; 5. Vivenza D et al. Int J Biol Markers, 2013; 6. Ahles TA, et al Psychooncology, 2003; 7.Small BJ et al. Cancer, 2011; 8. Collado-Hidalgo A et al. Brain Behav Immun, 2008; 9. Bower JE et al. J Clin Oncol, 2013; 10. Reinertsen KV et al. Brain Behav Immun, 2011; 11. Fernandez-de-las-Penas C et al. Breast Cancer Res Treat, 2012; 12. Hertz DL et al. Breast Cancer Res Treat, 2012; 13. Hertz DL et al. Ann Oncol, 2013; 14. Baldwin RM et al. Clin Cancer Res, 2012; 15. Hertz DL et al. Breast Cancer Res Treat. 2014; 16. Chang H et al. Ann Oncol, 2009; 17. Rizzo R et al. Breast Cancer Res Treat, 2010; 18. Sucheston LE et al. Breast Cancer Res Treat, 2011; Abraham JE et al. Clin Cancer Res 2014.

Cardiotoxicity ( Table 3 )

Five of the studies evaluated the genetic associations underlying cardiotoxicity[26], [27], [28], [29], and [30]. Trastuzumab and anthracyclines (epirubicin) were the main anti-cancer agents examined in the five studies. The human epidermal growth factor receptor 2 (HER2) SNPs were the most commonly investigated genetic variants. SNPs in the genes expressing the immunoglobulin receptors Fc fragments (FCGR), potassium channel, renin–angiotensin–aldosterone system and glutathione S-transferase were also examined.

Table 3 Genetic variants and cardiotoxicity.

First author (Year) Study objectives Inclusion criteria Exclusion criteria Sample size Measure(s) for the classification of cardiotoxicity Genes examined Association (Yes/No) Results Covariates controlled
(a) Genetic variants and cardiomyopathy – trastuzumab
Lemieux et al. [26]
  • Primary: To evaluate the associations betweenHER2polymorphisms (lle655Val, Ala1170Pro) and cardiomyopathy in breast cancer patients treated with trastuzumab
  • Secondary: To evaluate the association of cardiomyopathy with risk factors
  • (i) Non-metastaticHER2-positive invasive breast cancer
  • (ii) Diagnosed or completed chemotherapy between July 1, 2005 and January 1, 2010
  • (iii) Receiving trastuzumab in neo-adjuvant/ adjuvant setting
Nil Total: N = 237 (DNA and genotyping data was only obtained from 73 patients) Any of the following:
  • (i) Relative reduction from the baseline of more than 10% with a resulting left ventricular ejection fraction (LVEF) below 50% at follow up
  • (ii) Decline of the LVEF below 45%
  • Follow-up: not stated
Ile655Val Yes The Ile/Val genotype subgroup had higher risks of cardiomyopathy than the Ile/Ile genotype subgroup (OR = 5.87, 95% CI, p = 0.02)
  • (i) Baseline LVEF
  • (ii) Smoking status
Ala1170Pro No N.A.
Beauclair et al. [27] To examine the effect of the Ile665Val gene polymorphism on the pharmacodynamics of treatment by trastuzumab and tumor growth
  • (i) Advanced breast cancer
  • (ii) Overexpression of HER2
Pre-existing cardiovascular disease Total: N = 61
  • Relative reduction from the baseline by more than 20%
  • Follow-up (median): 22.4 months
Ile655Val Yes The Ile/Val genotype subgroup had a higher risk of cardiomyopathy (p = 0.0058)
All of the cases of cardiomyopathy were found in the Ile/Val genotype subgroup. No cases of cardiomyopathy were observed in the Ile/Ile and Val/Val genotype subgroups
Roca et al. [28] To investigate the predictive value of the lle655Val, FCGR2A and FCGR3A gene polymorphisms for cardiomyopathy and the efficacy of trastuzumab
  • (i) Node-positive, non-metastatic unilateral breast adenocarcinoma
  • (ii) HER2-positive breast tumor
  • (iii) Treated with adjuvant chemotherapy and trastuzumab
Nil Total: N = 132 Any of the following:
  • (i) Decline of the LVEF from the baseline, with a resulting LVEF below 50%
  • (ii) Relative reduction from the baseline of more than 15%
  • (iii) Discontinuation of trastuzumab in the case of cardiomyopathy or other clinical intolerances

Follow up (median): 46.8 months
lle655Val Yes The Val allele carriers experienced greater cardiovascular alterations (LVEF < 50%) than in individuals with the Ile/Ile genotype (69% vs 31% respectively; odds ratio 3.83; 95% CI [1.11–13.18]; p = 0.025)
(b) Genetic variants and cardiomyopathy – anthracycline (epirubicin)
Vivenzaet al. [30] To determine whether polymorphisms in the renin-angiotensin-aldosterone system and in the glutathione S-transferase family of phase 2 detoxification enzymes might be useful predictors of LVEF kinetics and the risk of developing congestive heart failure associated with adjuvant anthracycline chemotherapy
  • (i) Early breast cancer
  • (ii) Treated with adjuvant anthracycline chemotherapy
  • (iii) Underwent radical surgery
  • (iv) Adequate bone marrow, renal and hepatic function
  • (i) Distant metastases
  • (ii) Coronary heart disease, valvular heart disease
  • (i) Pre-treatment of left ventricular dysfunction
Total: N = 48 Any of the following:
  • (i) Overt congestive heart failure (grade III)
  • (ii) Decline of the LVEF below 50% (grade II) at any point during the 3 year follow-up

Follow up: 3 years
Genes in the renin-angiotensin-aldosterone system No N.A.
Genes in the glutathione S-transferase No N.A.
(c) Genetic variants and corrected QT prolongation
Kitagawaet al. [29] To investigate the association between corrected QT interval prolongation and adjuvant chemotherapy and the SNPs of the potassium channel genes
  • (i) Early stage breast cancer
  • (ii) Received FEC100 as adjuvant chemotherapy before or after surgery
  • (i) Major risk factors for corrected QT interval prolongation, such as hypokalemia, congestive heart failure or bradycardia
  • (ii) Receiving medication that potentially causes corrected QT interval prolongation
Total: N = 34 (27 patients agreed to genotyping) Graded according to the Common Terminology Criteria for Adverse Events KCNQ1 (rs757092) No N.A.
KCNH2 (rs3815459) No N.A.
KCNH2 (rs1805123) No N.A.
KCNH20 (rs3807375) No N.A.

Three of these five studies found a significant association between theHER2lle655Val polymorphism and trastuzumab-induced cardiotoxicity[26], [27], and [28]. Although Roca et al. reported that the presence of the Val allele in the lle/Val and Val/Val genotypes was associated with cardiotoxicity, it was observed that none of the patients with LVEF less than 50% were found in the Val/Val genotype group. In contrast, the two other studies only specified the lle/Val genotype as the associated genotype. No significant associations were found between other SNPs and cardiotoxicity.

Cognitive dysfunction ( Table 4 )

Two of the studies investigated the genetic variants underlying chemotherapy-induced cognitive dysfunction[31] and [32]. The apolipoprotein E (APOE) and catechol-O-methyltransferase (COMT) gene polymorphisms were examined in each study and positive associations were established for both genes.

Ahles et al. found that the ε4 allele of theAPOEgene was associated with poorer visual memory, spatial ability and psychomotor functioning in breast cancer and lymphoma survivors who received chemotherapy [31] . In another study conducted investigating the role ofCOMTgene polymorphisms on cognitive functions in breast cancer survivors, the authors found that the presence of the Val allele was associated with poorer attention, verbal fluency and motor speed [32] .

Table 4 Genetic variants and cognitive dysfunction.

First author (Year) Study objectives Inclusion criteria Exclusion criteria Sample size Measure(s) for the classification of cognitive dysfunction Genes examined Association (Yes/No) Results Covariates controlled
Ahleset al. [31] To compare the neuropsychological performance of long-term breast cancer and lymphoma treated with standard dose chemotherapy, in patients carrying the ε4 allele of the APOE gene to patients carrying other APOE alleles
  • (i) Long term survivors of breast cancer or lymphoma (disease-free)
  • (ii) Treated with chemotherapy
  • (iii) Completed a standardized battery of neuropsychological and psychological tests
  • (iv) At least five years post-diagnosis
  • (v) Receiving no cancer treatment, except tamoxifen
  • (vi) >18 years old when diagnosed
  • (vii) Fluent in English
  • (i) Central nervous system disease
  • (ii) Central nervous system radiation or intrathecal therapy
  • (iii) Neurobehavioral risk factors
Total: N = 80 Breast cancer: N = 51 Lymphoma: N = 29 Neuropsychological assessment battery, which evaluated:
  • - Attention
  • - Motor functioning
  • - Psychomotor functioning
  • - Spatial ability
  • - Verbal ability
  • - Verbal learning
  • - Verbal memory
  • - Visual memory
  • - Motor functioning
  • The subgroup containing the subjects carrying at least one ε4 allele demonstrated poorer visual memory (F(1, 78) = 4.91,p < 0.03)
  • The subgroup containing the subjects carrying at least one ε4 allele demonstrated poorer spatial ability (F(1, 78) = 3.91,p < 0.05)
  • The subgroup containing the subjects carrying at least one ε4 allele demonstrated poorer psychomotor functioning (F(1, 78) = 3.05,p < 0.08)
  • (i) Age
  • (ii) Gender
  • (iii) Education
  • (iv) Diagnosis
  • (v) WRAT-3 reading subset
Small et al. [32] To examine the moderating role of the COMT genotype on cognitive performance following treatment for breast cancer
  • (i) Stage 0 to II breast cancer survivors
  • (ii) >18 years of age
  • (iii) Fluent in English
  • (iv) Treated surgically with a lumpectomy or mastectomy
  • (v) Received chemotherapy and/or radiotherapy
  • (i) Presence of psychiatric or neurological disorders
  • (ii) History of cancer other than basal cell skin carcinoma
  • (iii) Chronic or life threatening diseases in which fatigue is a prominent symptom
  • Total:N = 334 Radiotherapy only (RT):N = 58
  • Chemotherapy only or chemotherapy and radiotherapy (CT):N = 72
  • Healthy controls (HC):N = 204
Neuropsychological assessment battery which evaluated:
  • - Attention
  • - Complex cognition
  • - Episodic memory
  • - Motor speed
  • - Overall cognition
  • - Verbal fluency
  • - Cognitive dysfunction: ⩾1.5 SD below the normative value
  • Cohort:
  • The Val allele carriers demonstrated poorer attention (mean = 52.19, SE = 0.45) than the Met/Met carriers (mean = 54.43, SE = 0.79) (p = 0.15,d = 0.31)
  • The Val allele carriers demonstrated poorer verbal fluency (mean = 47.35, SE = 0.77) than the Met/Met carriers (mean = 51.45, SE = 1.35) (p = 0.009,d = 0.34)
  • The Val allele carriers demonstrated poorer motor speed (mean = 46.37, SE = 0.78) than the individuals with the Met/Met genotype (mean = 49.76, SE = 1.38) (p = 0.033,d = 0.27)
  • Age
                Within each subgroup (CT/RT/HC):

Only the CT group yielded statistically significant results. The Val allele carriers demonstrated poorer attention than the individuals with the Met/Met genotype (p < 0.001, d = 0.96)

Fatigue ( Table 5 )

Four of the studies examined the genetic associations underlying fatigue[33], [34], [35], and [36]. TheCOMTgene and the genes expressing pro-inflammatory cytokines were investigated (IL1B-511,IL6-174,TNF-308,IL6Rand C-reactive protein).

Table 5 Genetic variants and fatigue.

First author (Year) Study objective Inclusion criteria Exclusion criteria Sample size Measure(s) for the classification of fatigue Genes examined Association (Yes/No) Results Covariates controlled
Collado-Hidalgo et al. [33] To test the hypothesis that variations in cytokine-related fatigue are influenced by genetic polymorphisms in the regulatory regions (promoters) of the genes that encode pro-inflammatory cytokines
  • (i) Survivors of breast cancer (stage 0, I, II)
  • (ii) Completed all treatment, with the exception of tamoxifen/aromatase inhibitor
  • (iii) One to five years post diagnosis
  • (i) Cancer recurrence
  • (ii) Chronic medical conditions involving the immune system or the regular use of immunosuppressive medications
Total: N = 47 Vitality subscale of SF-36:
  • (i) Fatigue: score ⩽55
  • (ii) Non-fatigue: score ⩾70
IL1B-511 No
  • The presence of at least one cytosine nucleotide in theIL1B-511gene was greater in the fatigued subgroup than the non-fatigued subgroup (95% CI = 0.91–16.6,p = 0.007)
  • After controlling for covariates, the association ofIL1B-511with fatigue remained significant but the association was not significant after controlling for depressive symptoms
  • After controlling for ethnicity (white vs. non-white), the association ofIL1B-511with fatigue remained significant
  • (i) Age
  • (ii) Depressive symptoms
  • (iii) Mastectomy
  • (iv) Breast reconstruction
  • (v) Breast irradiation
  • (vi) Ethnicity
  IL6-174 No
  • There was a greater representation of the G/G and C/C genotypes in the fatigue subgroup than in the non-fatigue subgroup (95% CI = 1.12-17.9,p = 0.27)
  • After controlling for covariates, the association ofIL6-174homozygosity with fatigue was insignificant
  • After controlling for ethnicity (white
  • vs. non-white), the association ofIL6-174homozygosity with fatigue remained significant
Bower et al. [34] Primary: To determine whether regulatory polymorphisms in three key pro-inflammatory cytokine genes (both individual and additive effects) were associated with fatigue in women who had recently completed treatment for early stage breast cancer
  • (i) Breast cancer (stage 0 to IIIA) patients
  • (ii) Aged 21–65
  • (iii) Completed primary cancer treatment within the past three months
  • (i) Endocrine therapy
  • (ii) Neurologic or immune-related medical conditions
  • (iii) Behavior known to influence the immune system
Total: N = 171
  • Multidimensional Fatigue Symptom Inventory: Higher scores indicated a higher level of fatigue
TNF-308 Yes
  • After controlling for covariates, the association between theTNF-308gene and fatigue remained significant (p = 0.034)
  • The G/G genotype subgroup had a higher level of fatigue than both the G/A and A/A genotype subgroups
  • (i) Age
  • (ii) BMI
  • (iii) Race/ethnicity
  • (iv) Treatment with chemotherapy
Secondary: To explore the association between pro-inflammatory cytokine genes and depression, memory complaints and sleep disturbance
IL6-174 Yes
  • After controlling for covariates, theIL6-174gene was associated with fatigue (p = 0.037)
  • The G/G genotype subgroup had a higher level of fatigue. The fatigue levels were slightly higher than those of the G/A subgroup and almost two times higher than those of the A/A subgroup
IL1B-511 No
  • The C/C genotype subgroup had a higher level of fatigue, but this was not statistically significant
Additive effect Yes
  • The assessment of the additive effect was based on the genetic risk score of high expression alleles in the three genes
  • After controlling for covariates, the genetic risk score was associated with higher levels of fatigue (p = 0.002)
Reinertsen et al. [35]
  • Primary: to explore the associations between fatigue and SNPs related to the inflammatory pathway in a large cohort of breast cancer survivors
  • Secondary: to explore the associations between the genetic variation in three selected genes and IL1B and IL6R mRNA expression in blood and serum hsCRP levels
  • (i) Survivors of breast cancer (stage II or III)
  • (ii) Aged ⩽ 75 in 2004/2005
  • (iii) Undergone loco-regional adjuvant radiotherapy, combined with adjuvant chemotherapy, with or without endocrine therapy
  • (i) Other cancer, except for basal cell carcinoma or carcinomain situof the uterine cervix
  • (ii) Recurrence of breast cancer
  • (iii) Depressive symptoms
  • Part 1: Total:N = 302
  • Part 2 (Follow up of part 1 after two to three years):
  • Total :N = 175
  • Note: The subjects in part 2 were from part 1
Fatigue Questionnaire (FQ):

Part 1:
  • - Chronic fatigue: score ⩾ 4 on FQ
  • - Non-fatigued: score < 4
  • - Part 2:- Persistent fatigue: score ⩾ 4 for both part 1 and 2 FQ assessments
  • - Never fatigued: score < 4 for both assessments
IL1B-511 No
  • N.A.
  • It is unclear if any variables were controlled in the analysis
  • After exclusion of the depressed subjects, the results remained non-significant
IL6-174 No
  • N.A.
  • N.A.
  • N.A.
Fernandez-de-las Penas et al. [36] To examine the influence of COMT Val158Met genotypes on cancer-related fatigue, post mastectomy pain and pressure pain hypersensitivity in breast cancer survivors
  • (i) Survivors of breast cancer (stage I - IIIA) without active cancer
  • (ii) Aged 25-65 years
  • (iii) Received a simple mastectomy or quadrantectomy, including those with breast reconstruction at the time of initial surgery or subsequent breast reconstruction
  • (iv) Finished co-adjuvant treatment, except endocrine therapy
  • (v) Experiencing neck or shoulder pain that began after the breast cancer surgery
  • (i) Receiving chemotherapy or radiotherapy at the time of the study
  • (ii) Breast surgery for cosmetic reasons or a prophylactic mastectomy
  • (iii) Medical inflammatory condition
  • (iv) Recurrent cancer
  • (v) Previous diagnosis of fibromyalgia syndrome
Total: N = 128 Piper fatigue scale:

A higher score indicated a higher level of fatigue
  • Higher levels of fatigue were recorded in the Met/Met and Val/Met genotype subgroups than in the Val/Val genotype subgroup (p < 0.01)
  • No significant differences were recorded in the fatigue scores between the subjects in the Val/Met and Met/Met genotype subgroups

The SNPs in the promoter regions of theIL1B-511andIL6-174genes were most frequently investigated. In the three studies that explored the associations betweenIL1B-511polymorphisms and fatigue, no significant association was established[33], [34], and [35]. Collado-Hidalgo et al. and Reinertsen et al. did not find an association betweenIL6-174polymorphisms and fatigue. However, Bower et al. found an association between the G/G genotype of theIL6-174polymorphism and fatigue. A similar association was reported forTNF-308polymorphisms, with higher levels of fatigue observed in individuals with G/G genotypes [34] . Bower et al. investigated the contributions of each SNP individually and examined the additive effect of the genes’ SNPs. The additive effects were calculated using the computed genetic risk scores of the high expression alleles in the three genes. A linear relationship between the genetic risk score and symptoms of fatigue was found: patients with more high-expression alleles reported significantly higher levels of fatigue [34] .

A positive association between fatigue and theCOMTVal158Met genotype was reported, as a higher incidence of fatigue was found to be correlated with the presence of the Met allele ofCOMT [36] .

Peripheral neuropathy ( Table 6 )

Most of the eight studies investigated the SNPs of the genes associated with paclitaxel metabolism (CYP2C8andCYP3A4/3A5), transport (ABCB1) and drug action (CYP1B1)[11], [12], [37], [38], [39], [40], [41], and [42]. The SNPs in the genes of the breast cancer type 1 (BRCA1) and Fanconi anemia group D2 (FANCD2) pathways and the ATP binding cassette transporter G1 gene (ABCG1) were also examined. All of the studies evaluated paclitaxel-induced peripheral neuropathy, except for one study, which evaluated peripheral neuropathy among patients treated with either paclitaxel or docetaxel[11], [12], [37], [38], [39], [40], and [41].

In the cytochrome genes (CYP2C8,CYP1B1andCYP3A4/3A5) investigated,CYP2C8gene polymorphisms were found to be associated with paclitaxel-induced peripheral neuropathy. Patients with theCYP2C8∗3allele have decreased paclitaxel metabolic activity and this leads to an increased in drug exposure. Three of the studies examined this association and found an association between theCYP2C8∗3allele and paclitaxel-induced peripheral neuropathy[11], [12], and [38]. The presence of eachCYP2C8∗3allele was found to double the risk of peripheral neuropathy. The association betweenCYP2C8gene polymorphisms and the occurrence of peripheral neuropathy was examined in a study of patients receiving paclitaxel and docetaxel [40] . However, an association was not established in this study.

Table 6 Genetic variants and peripheral neuropathy.

First author (Year) Study objective Inclusion criteria Exclusion criteria Sample size Measure(s) for the classification of peripheral neuropathy Genes examined Association (Yes/No) Results Covariates controlled
(a) Genetic variants and peripheral neuropathy – paclitaxel
Hertz et al. [11] To investigate the hypothesis that polymorphisms in the genes relevant to paclitaxel metabolism (CYP2C8 and CYP3A4/3A5), transport (ABCB1) or mechanism (CYP1B1) are associated with treatment efficacy and toxicity
  • (i) Breast cancer patients
  • (ii) Treated between 2005 and 2009
  • (iii) Received paclitaxel treatment guided by standard neoadjuvant protocols
Total: N = 111 (The data for only 109 patients were available for toxicity analysis) National Cancer Institute Common Toxicity Criteria for Adverse Events version 4.0

Neuropathy: ⩾grade 3+
CYP2C8 No The 3 allele carriers had a higher risk of neuropathy than the individuals with the 1/1 genotype (OR = 3.13, 95% CI: 0.89-11.01, p = 0.075)
CYP1B1 No N.A.
CYP3A4/3A5 No N.A.
Hertz et al. [12] To investigate the hypothesis that the increase in neuropathy risk associated with CYP2C8 can be replicated separately in independent cohorts of European-American and African-American breast cancer patients treated with paclitaxel
  • (i) Breast cancer patients
  • (ii) Treated between 2005 and 2011
  • (iii) Received neoadjuvant and/or adjuvant paclitaxel-containing regimens
Total: N = 411 National Cancer Institute Common Toxicity Criteria for Adverse Events version 4.0

Neuropathy: ⩾grade 2+
CYP2C8 Yes European-American subgroup

The 3/3 genotype subgroup had a higher risk of neuropathy. The lowest risk was in the 1/1 genotype subgroup (log rank p = 0.006)

Additive gene effect: the presence of each 3 allele (0, 1 or 2) doubled the neuropathy risk [HR (per allele) = 1.93, 95%, CI: 1.053.55, p = 0.032]

The association between the CYP2C8∗3 genotype and neuropathy risk remained significant after adjusting for covariates [HR (per allele) = 1.95, 95%, CI: 1.06–3.58, p = 0.031]

African-American subgroup

No 3/3 genotype individuals

The risk of neuropathy was highest in individuals with 1/3 genotype [HR (per allele) = 3.30, 95% CI: 1.04–10.45, p = 0.043]

Mixed-race cohort

There was a higher risk of neuropathy with the 3 allele [HR (per allele) = 1.98, 95% CI: 1.25–3.13, p = 0.004]

There was a higher risk of neuropathy in non-Europeans (HR = 1.76, 95% CI: 1.06–2.93, p = 0.030)
  • (i) Age
  • (ii) Prior diagnosis of diabetes
  • (iii) Taxane schedule
  • (iv) Use of neuropathy prophylaxis or treatment
  • (v) Self-reported race
Hertz et al. [38] To identify additional genetic variants within CYP2C8 and to individually assess thousands of SNPs that may influence paclitaxel pharmacokinetics
  • (i) Breast cancer patients
  • (ii) Treated between 2005 and 2011
  • (iii) Received neoadjuvant and/or adjuvant paclitaxel-containing regimens
Total: N = 412 National Cancer Institute Common Toxicity Criteria for Adverse Events version 4.0

Neuropathy: ⩾ grade 2+
CYP2C8 Yes The low metabolizer group had a significantly greater risk of neuropathy (HR = 1.722, 95 % CI 1.10–2.70, p = 0.018)

Using the likelihood ratio test, the influences of the 2 and 4 SNPs were not independently significant (2: p = 0.847, 4:

p = 0.408), whereas the 3 variant was significantly associated with neuropathy (p = 0.03)
ABCG1 Yes The intronic SNP in ABCG1 (rs492338) exceeded the exploratory significance threshold of 0.001 for an association with neuropathy in the Caucasian cohort [HR (per allele) = 2.11, 95% CI 1.36 to 3.29, p = 0.0008]

However, this was not significant in the non-Caucasian replication group (p = 0.54)
Chang et al. [39] To evaluate the association between the clinical outcome (safety and efficacy) of paclitaxel monotherapy in metastatic breast cancer patients and the ABCB1 gene polymorphisms 2677 or 3435
  • (i) Histologically confirmed breast cancer, receiving paclitaxel monotherapy
  • (ii) Not pregnant
  • (iii) Age ⩾ 18 years
  • (iv) ECOG performance status of 0-2
  • (v) A life expectancy of > 12 weeks
  • (vi) Clinically or radiographically measurable disease
  • (vii) Adequate renal, hepatic, and bone marrow function
  • (i) Prior taxane therapy
  • (ii) Prior chemotherapy, completed within four weeks prior to entry into the study
  • (iii) History of other neoplasm
  • (iv) History of ventricular arrhythmias or congestive heart failure
  • (v) History of pre-existing motor or sensory neuropathy ⩾grade 1
Total: N = 121 (108 patients agreed to the examination of their ABCB1 genotype)

Note: gDNA could not be extracted for some patients
Neuropathy graded using the National Cancer Institute Common Toxicity Criteria, version 3.0 ABCB1 2677 No N.A.
ABCB1 3435 No N.A.  
Sucheston et al. [41]
  • (i) To evaluate the associations of theBRCA1andFANCD2SNPs with paclitaxel-induced toxicities
  • (ii) To examine the gene expression levels ofBRCA1andFANCD2
  • (i) Node positive or high-risk (tumor ⩾2 cm) node-negative operable stage II or III invasive breast cancer with known estrogen and progesterone receptor statuses
  • (ii) Undergone breast surgery
  • (i) Prior cancer
  • (ii) Prior chemotherapy
Total: N = 905 (DNA and genotyping data was only obtained from 888 patients)
  • (i) Neuropathy graded using the National Cancer Institute Common Toxicity Criteria, version 3.0
  • (ii) Functional Assessment of Cancer Therapy – Taxane
Genes in the BRCA1 pathway No N.A.
  • (i) Age
  • (ii) European ancestry
  • (iii) Treatment group
Genes in the FANCD2 pathway Yes There was an increase in the odds of peripheral neuropathy (using the CTC criteria) with the SNPs:
  • - rs7648104: 1.86 (95% CI = 1.30,2.65)
  • - rs7637888: 1.87 (95% CI = 1.30, 2.67)
  • - rs6786638: 1.90 (95% CI = 1.32, 2.72)
  • - rs6442150: 1.89 (95% CI = 1.32,2.72)
  • - (p < 0.001)
Abraham et al. [42] To evaluate the association between previously investigated SNPs and taxane-related sensory neuropathy (TRSN), using genotype data from a study of chemotherapy-related toxicity in patients with breast cancer
  • (i) Breast cancer patients treated with paclitaxel from the tAnGo and Neo-tango trials
Total: N = 1335
  • (i) Neuropathy graded using the National Cancer Institute Common Toxicity Criteria, version 2.0
CYP2C8 Yes 4 variant is associated with risk of TRSN. OR, 1.48; 95% CI, 1.02–2.15; p = 0.04
  • (i) Body mass index
  • (ii) Trial status
ABCB1(rs2032582) Yes OR, 1.22; 95% CI, 1.03–1.45; p = 0.02
ABCB1(rs3213619) Yes OR, 0.47; 95% CI, 0.28–0.79; p = 0.004
ABCC2(rs8187710) Yes OR, 0.63; 95% CI, 0.42–0.93; p = 0.02
CYP1B1 Yes 3 variant is associated with risk of TRSN. OR, 0.81; 95% CI, 0.68–0.96; p = 0.02
TUBB2A(rs9501929) Yes OR, 1.80; 95% CI, 1.20–2.72; p = 0.005
KIAA0146-PRKD(rs6473187) Yes OR, 1.48; 95% CI, 1.01–2.17; p = 0.04
SLCO1B1(rs3829306) Yes OR, 0.66; 95% CI, 0.44–1.01; p = 0.05
EPHA6(rs301927) Yes OR, 1.35; 95% CI, 1.07–1.70; p = 0.01
(b) Genetic variants and peripheral neuropathy – paclitaxel and docetaxel
Rizzo et al. [40] To evaluate whether the known variant alleles of the CYP2C8, ABCB1 and CYP1B1 genes are associated with Taxane-induced toxicity in adult Caucasian breast cancer patients
  • (i) Histologically confirmed diagnosis of breast cancer
  • (ii) Treatment with docetaxel or paclitaxel as adjuvant, metastatic or neo-adjuvant therapy
Total: N = 95 (86 patients were evaluated for neurotoxicity) National Cancer Institute Common Toxicity Criteria, version 3.0 CYP2C8 No N.A.
CYP1B1 No N.A.
(c) Genome-Wide association study and peripheral neuropathy onset and severity – paclitaxel
Baldwin et al. [37] To identify the genetic predictors of the onset and severity of sensory peripheral neuropathy in paclitaxel-treated women (monotherapy)
  • (i) Histologically confirmed invasive carcinoma of the breast
  • (ii) 0 to 3 axillary nodes positive for cancer
  • (iii) ECOG performance status of 0 to 1
  • (iv) Disease-free from any prior malignancies for at least five years
  • (v) Absence of congestive heart failure or myocardial infarction in the past six months
  • (i) Locally advanced, inflammatory or metastatic breast cancer or involvement of dermal lymphatics
  • (ii) Undergone previous trastuzumab, anthracycline, chemotherapy or hormonal therapy (except tamoxifen or other selective estrogen receptor modulators for prevention or other indications)
  • Discovery set: Total:N = 855
  • European Replication: Total:N = 154
  • African American Replication: Total:N = 117
National Cancer Institute Common Toxicity Criteria for Adverse Events version 2.0

The timing of sensory peripheral neuropathy was assessed with a time-to-event approach, in which the event was defined as the first incidence of neuropathy ⩾grade 2
For the onset of sensory peripheral neuropathy, no SNPs achieved genome-wide significance. Seven SNPs achieved marginal significance. SNPs in the FGD4 gene (rs10771973) and EPHA5 gene (rs7349683) were biologically relevant No (did not achieve genome-wide significance) FGD4 (rs10771973): Both the discovery set and replication cohorts experienced a greater onset of neuropathy

Discovery cohort:

HR, 1.57; 95% CI, 1.30–1.91; p = 2.6 × 10−6

European replication:

HR, 1.72; 95% CI, 1.06–2.80; p = 0.013

African American replication:

HR, 1.93; 95% CI, 1.13–3.28; p = 0.0067.

EPHA5 (rs7349683):

Only the discovery set experienced a greater onset of neuropathy

Discovery cohort: HR, 1.63; 95% CI, 1.34-1.98; p = 9.6 × 10−7
For the severity of sensory peripheral neuropathy, a SNP in the FZD3 gene (rs7001034) achieved genome-wide significance Yes (achieved genome-wide significance) FZD3 (rs7001034):
Only the discovery set demonstrated the most severe neuropathy
OR = 0.57; 95% CI, 0.48–0.69; p = 3.1 × 10−9

The SNPs of theABCB1,BRCA1andFANCD2genes were also examined. FourFANCD2SNPs were reported to be significantly associated with the occurrence of peripheral neuropathy, specifically paclitaxel-induced peripheral neuropathy [41] . Three of the studies reported that there was no association betweenABCB1SNPs and the occurrence of peripheral neuropathy[11], [39], and [40]. No association was found betweenBRCA1SNPs and peripheral neuropathy. One study reported an association between paclitaxel-induced peripheral neuropathy and theABCG1SNP, but this association was not reproducible in the replication cohort [38] .

A genome-wide association study was conducted to analyze the association of selective SNPs with the onset and severity of paclitaxel-induced peripheral neuropathy [37] . Genome-wide significance was reported when the SNPs were analyzed for their association with the severity of peripheral neuropathy. A SNP inFZD3(rs7001034) was reported to be significantly associated with more severe peripheral neuropathy. However, genome-wide significance was not established when the SNPs were analyzed for their associations with the onset of peripheral neuropathy.

In a replication study conducted by Abraham et al. to investigate the roles of 70 SNPs in 50 genes for their association with the risk of peripheral neuropathy, a total of nine uncorrelated SNPs were found to be significantly associated with peripheral neuropathy [42] . Among the nine SNPs, the SNPs of the ABCB1 (rs3213619) and TUBB2A (rs9501929) genes had the strongest association with peripheral neuropathy. CYP2C8∗3 variant was not found to be associated with peripheral neuropathy. Instead, the CYP2C8∗4 variant showed an increased risk of neuropathy.


In this systematic review, positive associations between genetic variants and chronic and late toxicities of breast cancer treatment were established in more than half of the studies. Among the four toxicities included in this review, peripheral neuropathy has the known highest number of associated genes reported. While certain genetic variants such as the CYP2C8∗3 variant and theHER2lle655Val polymorphism were found to have significant association, it can be observed that the associations were not consistently replicated across the reviewed studies. However, it must be noted that conflicting and negative results do not necessarily undermine the contribution of genetics to the occurrence of these toxicities. The presence of methodological differences and flaws may account for the inconsistencies observed in the reported results, which may limit the comparability of the studies’ results.

The use of different study designs could explain the differences observed[43] and [44]. For example, three of the studies investigated the genetic association between theIL6-174SNP and fatigue. Two of these studies used a typical case-control study design, whereas the third study used a quantitative trait association study design, examining the direct association between theIL6-174SNP and the levels of fatigue[33], [34], [35], and [45]. A case-control study uses a discrete phenotypic trait to segregate the study population into cases and controls, whereas a quantitative trait association study aims to establish the association between the gene of interest and a continuous trait. The two study designs possess distinctive advantages and disadvantages. Most importantly, different statistical methods for data analysis are required, due to the nature of the trait that is unique to each study design, and these differences may yield different results [45] .

The inclusion of study participants who received different types of chemotherapy agent may have contributed to the contradictory findings. In the studies conducted by Hertz et al., all of the patients were administered paclitaxel, whereas in the study conducted by Rizzo et al., the minority of the patients received paclitaxel and the majority of the patients (74%) received docetaxel[11], [12], and [40]. Although the structural and mechanistic properties of docetaxel and paclitaxel are similar, it must be emphasized that these drugs are metabolized differently. The metabolism of paclitaxel is dependent on both the CYP2C8 and CYP3A4 enzyme systems, whereas docetaxel is mainly metabolized by CYP3A4 [46] . Systematic concomitant treatment with another neurotoxic agent that does not interact with the CYP2C8 enzyme system could reduce the apparent effect ofCYP2C8SNPs on the neuropathy risk, possibly leading to a false-negative finding.

The lack of a clear phenotype to distinguish between cases and controls may also threaten the results of genetic association studies[43] and [44]. Although Ahles et al. and Small et al. observed poorer cognitive function in the ε4 allele and Val allele carriers than in the non-carriers, the degree of cognitive decline reported by both groups was either subtle in nature, or lacked an important difference when the treatment arms were compared to the control arms[31] and [32]. As the degree of cognitive change observed in both studies may not be clinically relevant, it may be difficult to establish the true phenotype, which is critical for a genetic association study. The use of different definitions and methods to identify the phenotypes may have compounded the issue of poor comparability across the studies. Guidelines are available to harmonize the methodological differences between studies that are designed to evaluate chronic and late toxicities[47] and [48]. However, these recommendations may not be universally applicable to all studies. As such, it is appropriate for studies to use the tools that have been validated for their population of interest and that are sufficiently sensitive to differentiate between the phenotypes in the cases and controls. It is also important for studies to provide detailed descriptions and justifications for the definitions used in the identification of the phenotype of interest. Failing to adjust for clinical, behavioral or environmental factors may also have affected the phenotype, independently of the gene of interest, which in turn may have affected the precision of the results [49] . For example, as depression can affect the severity of cancer-related fatigue, it necessary for studies to take into account the confounding effect of depression in their analysis and study design[50] and [51].

A robust sample size is one of the key determinants for quality genetic association studies; a genetic association study must be sufficiently powered to detect associations[43], [44], [52], and [53]. The majority of the studies did not indicate the sample size calculation in their statistical analysis. Most of the studies in all four categories of toxicities had small sample sizes, which may have led to underpowered studies. Of the 19 studies, 6 had a total study cohort of less than 100 participants. For rare genotypes, such as the Val/Val genotype of theHER2lle655Val polymorphism, the resultant small subgroups may have resulted in non-valid inferences about the associations[27] and [54]. Moreover, when small sample sizes are used, limited data on toxicity is generated for analysis, as only a few toxicity cases will be observed within the sample. For instance, Rizzo et al. only recorded peripheral neuropathy in seven out of eighty-six (8%) study participants. Given so few cases of peripheral neuropathy, the conclusions derived may require confirmation by larger studies. Another consideration is the ethnic background of the study population, leading to a different frequency of occurrence of the SNPs being investigated. Since SNPs are germline mutation, the distribution may differ from one population to another and this could lead to differences in terms of associating the significance of the chronic toxicities.

A number of the studies did not adequately justify that the manifestation of the toxicities observed was due to the anti-cancer treatment. Only 9 of the 19 studies either included a healthy control group or performed a baseline measure to serve as a comparison. It is unclear if the associations reported in the studies that included neither baseline options were specific to the breast cancer treatment. It is possible that the associated genotype may also have conferred a risk to the non-cancer populations. It is also important that studies report the methodological details of how a data set is produced and its overall reproducibility to allow critical assessment of the results and valid comparisons across studies.

The lack of transparency in the reporting of methods and results in genetic association studies makes the interpretation of their results and the assessment of the internal and external validity difficult. The STREGA guidelines are a set of guiding principles to maximize the transparency, quality and completeness of reporting in genetic association studies. It was evident that most of the studies reviewed did not meet the reporting criteria recommended by STREGA. Of the five main categories of information recommended for inclusion by STREGA, the reporting of the genotyping methods and errors did not meet the stringent criteria. Most of the studies did not adequately report the genotyping methods used, especially the error and call rates, nor did they report whether the genotyping was performed in batches or simultaneously. It is important to quantify error and call rates as they may have a significant effect on the ability to detect linkage or association[52], [55], and [56]. Evidence has also suggested that genotyping in batches may introduce an undesirable batch effect into an analysis, as the composition and size of the batch can generate a 2 to 3% discordance in the results [57] . In a typical case-control genetic association study, population stratification exists when both the cases and controls are recruited from multi-ethnic groups whose allele frequencies differ for the investigated gene variants[43], [44], and [58]. This inadvertently introduces a confounding effect into the analysis and leads to false positive associations. Studies must therefore report whether population stratification is assessed and adjusted during analysis.

Treatment-associated toxicities can compromise treatment outcomes through dose attenuation and/or treatment discontinuation. In survivors, the manifestation of undesirable chronic and late toxicities can adversely affect patients’ quality of life and their ability to perform their social roles efficiently. However, the sporadic and unpredictable nature of chronic and late toxicities suggests a potential role for genetics in their etiologies. A patient’s germline genome plays an important role in the determination of their individual response to therapeutic agents[59] and [60]. Thus, the presence of germline mutations in genes that encode for drug-metabolizing enzymes, transporters, receptors and signaling pathways may result in the large disparity in individual responses to anti-cancer treatments. These mutations may contribute significant heterogeneity to the occurrence of undesirable treatment-related toxicities. Most of the genes in the genetic association studies were evaluated based on their biological plausibility with the outcome of interest. Establishing a significant association between the genes of interest and the relevant toxicities could provide greater insights into the mechanisms underlying these toxicities at a molecular level. However, the use of genetic information to elucidate the mechanisms behind these toxicities poses certain challenges. For example, recent work has suggested that the dysregulation of pro-inflammatory cytokines levels, through the use of anti-cancer treatments, may provoke a neuroimmunoendocrine response that results in behavioral symptoms such as anxiety, fatigue and cognitive dysfunction[61], [62], [63], [64], [65], and [66]. Although studies have identified several genetic polymorphisms in the cytokine and cytokine receptor genes that can alter gene expression or protein structure, the functionality of certain genetic variants remains poorly understood [18] . It is only through functional genetic variants that genuine disease associations can truly be established. A better understanding of these genetic variants can be achieved using well-designed genetic association studies with large sample sizes and, most importantly, reproducible results

One limitation of this systematic review is that the review focuses primarily on the four specific chronic and late toxicities. Hence, the genetic associations underlying endocrine treatment-induced toxicities such as tamoxifen-induced hot flashes and aromatase inhibitor-associated arthralgia are not included. Nevertheless, it must be noted that the genetics variants associated with these toxicities are extensively described in recent reviews[22] and [23]. In addition, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist was not used in this review. The PRISMA statement have provided extensive guidance on how to perform a systematic review [67] . However, their focus is mainly on interventional studies (especially for randomized controlled trials). As such, certain elements that are highlighted in the PRISMA checklist may not be applicable for this systematic review which focuses mainly on genetic association studies. In this review, the STREGA guidelines were used to assess the quality of reporting of information in genetic association studies. Through assessing the quality of reporting of individual study, readers are able to identify the potential biases that may threaten the validity of the findings.

This systematic review has successfully reported and evaluated the findings of published studies that have investigated the genetic associations between cancer treatment and chronic/late toxicities in the breast cancer population. While current evidence has suggested that certain genetic variants are associated with the four toxicities, a number of methodological concerns in the reviewed studies may limit the interpretation and comparability of the findings. This is further compounded by the lack of transparency in the reporting of information since most studies did not conform to the STREGA guidelines. As such, more well-designed genetic association studies are warranted to confirm and validate these findings.

Future studies would benefit by addressing key methodological issues such as small sample size and population stratification, and improving the quality of data reporting by incorporating robust criteria. A concerted effort is also required to harmonize the methodological differences across various studies. It is timely to address these crucial issues, in order to ensure that that the findings from genetic association studies can be more effective to influence the development of personalized therapeutics for addressing toxicities issues in cancer patients and survivors.

Conflict of Interest

The authors have declared no conflict of interest.


This study was financed by research grants awarded by the National University of Singapore (R-148-000-166-112), the National Cancer Centre Singapore (NRFCB12131) and the National Medical Research Council Singapore (NMRC/CIRG/1386/2014).


  • [1] American Cancer Society. Cancer treatment and survivorship facts & figures 2012–2013. (American Cancer Society, Atlanta, 2012)
  • [2] D. Stan, C.L. Loprinzi, K.J. Ruddy. Breast cancer survivorship issues. Hematol Oncol Clin North Am. 2013;27:805-827 ix Crossref
  • [3] C.E. DeSantis, C.C. Lin, A.B. Mariotto, R.L. Siegel, K.D. Stein, J.L. Kramer, et al. Cancer treatment and survivorship statistics, 2014. CA Cancer J Clin. 2014;64:252-271
  • [4] D. Cardinale, A. Colombo, G. Lamantia, N. Colombo, M. Civelli, G. De Giacomi, et al. Anthracycline-induced cardiomyopathy: clinical relevance and response to pharmacologic therapy. J Am Coll Cardiol. 2010;55:213-220 Crossref
  • [5] M.I. Gharib, A.K. Burnett. Chemotherapy-induced cardiotoxicity: current practice and prospects of prophylaxis. Eur J Heart Fail. 2002;4:235-242 Crossref
  • [6] T.J. Postma, J.B. Vermorken, A.J. Liefting, H.M. Pinedo, J.J. Heimans. Paclitaxel-induced neuropathy. Ann Oncol. 1995;6:489-494
  • [7] M. Hofman, J.L. Ryan, C.D. Figueroa-Moseley, P. Jean-Pierre, G.R. Morrow. Cancer-related fatigue: the scale of the problem. Oncologist. 2007;12(Suppl. 1):4-10
  • [8] J. Vardy. Cognitive function in breast cancer survivors. Cancer Treat Res. 2009;151:387-419 Crossref
  • [9] C. Anderson-Hanley, M.L. Sherman, R. Riggs, V.B. Agocha, B.E. Compas. Neuropsychological effects of treatments for adults with cancer: a meta-analysis and review of the literature. J Int Neuropsychol Soc. 2003;9:967-982
  • [10] J.S. Myers. Chemotherapy-related cognitive impairment: the breast cancer experience. Oncol Nurs Forum. 2012;39:E31-E40 Crossref
  • [11] D.L. Hertz, A.A. Motsinger-Reif, A. Drobish, S.J. Winham, H.L. McLeod, L.A. Carey, et al. CYP2C8∗3 predicts benefit/risk profile in breast cancer patients receiving neoadjuvant paclitaxel. Breast Cancer Res Treat. 2012;134:401-410 Crossref
  • [12] D.L. Hertz, S. Roy, A.A. Motsinger-Reif, A. Drobish, L.S. Clark, H.L. McLeod, et al. CYP2C8∗3 increases risk of neuropathy in breast cancer patients treated with paclitaxel. Ann Oncol. 2013;24:1472-1478 Crossref
  • [13] S.B. Park, D. Goldstein, A.V. Krishnan, C.S. Lin, M.L. Friedlander, J. Cassidy, et al. Chemotherapy-induced peripheral neurotoxicity: a critical analysis. CA Cancer J Clin. 2013;63:419-437 Crossref
  • [14] K. Osmani, S. Vignes, M. Aissi, F. Wade, P. Milani, B.I. Lévy, et al. Taxane-induced peripheral neuropathy has good long-term prognosis: a 1- to 13-year evaluation. J Neurol. 2012;259:1936-1943 Crossref
  • [15] K. Shimozuma, Y. Ohashi, A. Takeuchi, T. Aranishi, S. Morita, K. Kuroi, et al. Taxane-induced peripheral neuropathy and health-related quality of life in postoperative breast cancer patients undergoing adjuvant chemotherapy: N-SAS BC 02, a randomized clinical trial. Support Care Cancer. 2012;20:3355-3364 Crossref
  • [16] J.A. Boyette-Davis, J.P. Cata, L.C. Driver, D.M. Novy, B.M. Bruel, D.L. Mooring, et al. Persistent chemoneuropathy in patients receiving the plant alkaloids paclitaxel and vincristine. Cancer Chemother Pharmacol. 2013;71:619-626 Crossref
  • [17] L.J. Leandro-Garcia, L. Inglada-Perez, G. Pita, E. Hjerpe, S. Leskela, C. Jara, et al. Genome-wide association study identifies ephrin type A receptors implicated in paclitaxel induced peripheral sensory neuropathy. J Med Genet. 2013;50:599-605 Crossref
  • [18] A.J. Smith, S.E. Humphries. Cytokine and cytokine receptor gene polymorphisms and their functionality. Cytokine Growth Factor Rev. 2009;20:43-59 Crossref
  • [19] C.C. Reyes-Gibby, X. Wu, M. Spitz, R. Kurzrock, M. Fisch, E. Bruera, et al. Molecular epidemiology, cancer-related symptoms, and cytokines pathway. Lancet Oncol.. 2008;9:777-785 Crossref
  • [20] S. Marsh, H.L. McLeod. Pharmacogenetics and oncology treatment for breast cancer. Expert Opin Pharmacother. 2007;8:119-127 Crossref
  • [21] M. Pirmohamed. Pharmacogenetics: past, present and future. Drug Discov Today. 2011;16:852-861 Crossref
  • [22] M. Del Re, A. Michelucci, P. Simi, R. Danesi. Pharmacogenetics of anti-estrogen treatment of breast cancer. Cancer Treat Rev. 2012;38:442-450 Crossref
  • [23] M.J. Higgins, V. Stearns. Pharmacogenetics of endocrine therapy for breast cancer. Annu Rev Med. 2011;62:281-293 Crossref
  • [24] E. von Elm, D. Moher, J. Little. Reporting genetic association studies: the STREGA statement. Lancet. 2009;374:98-100 Crossref
  • [25] J. Little, J.P. Higgins, J.P. Ioannidis, D. Moher, F. Gagnon, E. von Elm, et al. STrengthening the REporting of Genetic Association Studies (STREGA)—an extension of the STROBE statement. Genet Epidemiol. 2009;33:581-598 Crossref
  • [26] J. Lemieux, C. Diorio, M.A. Côté, L. Provencher, F. Barabé, S. Jacob, et al. Alcohol and HER2 polymorphisms as risk factor for cardiotoxicity in breast cancer treated with trastuzumab. Anticancer Res. 2013;33:2569-2576
  • [27] S. Beauclair, P. Formento, J.L. Fischel, W. Lescaut, R. Largillier, E. Chamorey, et al. Role of the HER2 [Ile655Val] genetic polymorphism in tumorogenesis and in the risk of trastuzumab-related cardiotoxicity. Ann Oncol. 2007;18:1335-1341 Crossref
  • [28] L. Roca, V. Dieras, H. Roche, E. Lappartient, P. Kerbrat, L. Cany, et al. Correlation of HER2, FCGR2A, and FCGR3A gene polymorphisms with trastuzumab related cardiac toxicity and efficacy in a subgroup of patients from UNICANCER-PACS 04 trial. Breast Cancer Res Treat. 2013;139:789-800 Crossref
  • [29] K. Kitagawa, K. Kawada, S. Morita, M. Inada, A. Mitsuma, M. Sawaki, et al. Prospective evaluation of corrected QT intervals and arrhythmias after exposure to epirubicin, cyclophosphamide, and 5-fluorouracil in women with breast cancer. Ann Oncol. 2012;23:743-747 Crossref
  • [30] D. Vivenza, M. Feola, O. Garrone, M. Monteverde, M. Merlano, C. Lo Nigro. Role of the renin–angiotensin–aldosterone system and the glutathione S-transferase Mu, Pi and Theta gene polymorphisms in cardiotoxicity after anthracycline chemotherapy for breast carcinoma. Int. J Biol. 2013;28:e336-e347
  • [31] T.A. Ahles, A.J. Saykin, W.W. Noll, C.T. Furstenberg, S. Guerin, B. Cole, et al. The relationship of APOE genotype to neuropsychological performance in long-term cancer survivors treated with standard dose chemotherapy. Psychooncology. 2003;12:612-619 Crossref
  • [32] B.J. Small, K.S. Rawson, E. Walsh, H.S. Jim, T.F. Hughes, L. Iser, et al. Catechol-O-methyltransferase genotype modulates cancer treatment-related cognitive deficits in breast cancer survivors. Cancer. 2011;117:1369-1376 Crossref
  • [33] A. Collado-Hidalgo, J.E. Bower, P.A. Ganz, M.R. Irwin, S.W. Cole. Cytokine gene polymorphisms and fatigue in breast cancer survivors: early findings. Brain Behav Immun. 2008;22:1197-1200 Crossref
  • [34] J.E. Bower, P.A. Ganz, M.R. Irwin, S. Castellon, J. Arevalo, S.W. Cole. Cytokine genetic variations and fatigue among patients with breast cancer. J Clin Oncol. 2013;31:1656-1661 Crossref
  • [35] K.V. Reinertsen, G.I. Grenaker Alnaes, H. Landmark-Hoyvik, J.H. Loge, E. Wist, V.N. Kristensen, et al. Fatigued breast cancer survivors and gene polymorphisms in the inflammatory pathway. Brain Behav Immun. 2011;25:1376-1383 Crossref
  • [36] C. Fernandez-de-las-Penas, C. Fernandez-Lao, I. Cantarero-Villanueva, S. Ambite-Quesada, I. Rivas-Martinez, R. del Moral-Avila, et al. Catechol-O-methyltransferase genotype (Val158met) modulates cancer-related fatigue and pain sensitivity in breast cancer survivors. Breast Cancer Res Treat. 2012;133:405-412 Crossref
  • [37] R.M. Baldwin, K. Owzar, H. Zembutsu, A. Chhibber, M. Kubo, C. Jiang, et al. A genome-wide association study identifies novel loci for paclitaxel-induced sensory peripheral neuropathy in CALGB 40101. Clin Cancer Res. 2012;18:5099-5109 Crossref
  • [38] D.L. Hertz, S. Roy, J. Jack, A.A. Motsinger-Reif, A. Drobish, L.S. Clark, et al. Genetic heterogeneity beyond CYP2C8∗3 does not explain differential sensitivity to paclitaxel-induced neuropathy. Breast Cancer Res Treat. 2014;45:245-254
  • [39] H. Chang, S.Y. Rha, H.C. Jeung, C.K. Im, J.B. Ahn, W.S. Kwon, et al. Association of the ABCB1 gene polymorphisms 2677G>T/A and 3435C>T with clinical outcomes of paclitaxel monotherapy in metastatic breast cancer patients. Ann Oncol. 2009;20:272-277
  • [40] R. Rizzo, F. Spaggiari, M. Indelli, G. Lelli, O.R. Baricordi, P. Rimessi, et al. Association of CYP1B1 with hypersensitivity induced by taxane therapy in breast cancer patients. Breast Cancer Res Treat. 2010;124:593-598 Crossref
  • [41] L.E. Sucheston, H. Zhao, S. Yao, G. Zirpoli, S. Liu, W.E. Barlow, et al. Genetic predictors of taxane-induced neurotoxicity in a SWOG phase III intergroup adjuvant breast cancer treatment trial (S0221). Breast Cancer Res Treat. 2011;130:993-1002 Crossref
  • [42] J.E. Abraham, Q. Guo, L. Dorling, J. Tyrer, S. Ingle, R. Hardy, et al. Replication of genetic polymorphisms reported to be associated with taxane-related sensory neuropathy in patients with early breast cancer treated with Paclitaxel. Clin Cancer Res. 2014;20:2466-2475 Crossref
  • [43] A. Li, D. Meyre. Challenges in reproducibility of genetic association studies: lessons learned from the obesity field. Int J Obes (Lond). 2013;37:559-567 Crossref
  • [44] H.J. Cordell, D.G. Clayton. Genetic association studies. Lancet. 2005;366:1121-1131 Crossref
  • [45] J.T. Dudley, K.J. Karczewski. Exploring personal genomics. (Oxford University Press, Oxford, 2013)
  • [46] V.A. de Weger, J.H. Beijnen, J.H. Schellens. Cellular and clinical pharmacology of the taxanes docetaxel and paclitaxel – a review. Anticancer Drugs. 2014;25:488-494 Crossref
  • [47] J.S. Wefel, J. Vardy, T. Ahles, S.B. Schagen. International Cognition and Cancer Task Force recommendations to harmonise studies of cognitive function in patients with cancer. Lancet Oncol. 2011;12:703-708 Crossref
  • [48] A. Seidman, C. Hudis, M.K. Pierri, S. Shak, V. Paton, M. Ashby, et al. Cardiac dysfunction in the trastuzumab clinical trials experience. J Clin Oncol. 2002;20:1215-1221 Crossref
  • [49] K.L. Lunetta. Genetic association studies. Circulation. 2008;118:96-101 Crossref
  • [50] K. Demyttenaere, J. De Fruyt, S.M. Stahl. The many faces of fatigue in major depressive disorder. Int J Neuropsychopharmacol. 2005;8:93-105 Crossref
  • [51] D.S. Baldwin, G.I. Papakostas. Symptoms of fatigue and sleepiness in major depressive disorder. J Clin Psychiatry. 2006;67(Suppl. 6):9-15
  • [52] A.T. Hattersley, M.I. McCarthy. What makes a good genetic association study?. Lancet. 2005;366:1315-1323 Crossref
  • [53] P.R. Burton, A.L. Hansell, I. Fortier, T.A. Manolio, M.J. Khoury, J. Little, et al. Size matters: just how big is BIG?: Quantifying realistic sample size requirements for human genome epidemiology. Int J Epidemiol. 2009;38:263-273 Crossref
  • [54] K.Y. Chan, A.N. Cheung, S.P. Yip, H.H. Ko, T.W. Lai, U.S. Khoo. Population-based case-control study of HER2 genetic polymorphism and breast cancer risk. J Natl Cancer Inst. 2002;94:1581-1582 Crossref
  • [55] E. Sobel, J.C. Papp, K. Lange. Detection and integration of genotyping errors in statistical genetics. Am J Hum Genet. 2002;70:496-508 Crossref
  • [56] F. Pompanon, A. Bonin, E. Bellemain, P. Taberlet. Genotyping errors: causes, consequences and solutions. Nat Rev Genet. 2005;6:847-859
  • [57] K. Miclaus, R. Wolfinger, S. Vega, M. Chierici, C. Furlanello, C. Lambert, et al. Batch effects in the BRLMM genotype calling algorithm influence GWAS results for the Affymetrix 500K array. Pharmacogenomics J. 2010;10:336-346 Crossref
  • [58] D.G. Clayton, N.M. Walker, D.J. Smyth, R. Pask, J.D. Cooper, L.M. Maier, et al. Population structure, differential bias and genomic control in a large-scale, case-control association study. Nat Genet. 2005;37:1243-1246 Crossref
  • [59] D.L. Hertz, H.L. McLeod. Use of pharmacogenetics for predicting cancer prognosis and treatment exposure, response and toxicity. J Hum Genet. 2013;58:346-352 Crossref
  • [60] W.P. Yong, F. Innocenti, M.J. Ratain. The role of pharmacogenetics in cancer therapeutics. Br J Clin Pharmacol. 2006;62:35-46 Crossref
  • [61] J.E. Bower, D.M. Lamkin. Inflammation and cancer-related fatigue: mechanisms, contributing factors, and treatment implications. Brain Behav Immun. 2013;30(Suppl):S48-S57 Crossref
  • [62] J.S. Myers. The possible role of cytokines in chemotherapy-induced cognitive deficits. Adv Exp Med Biol. 2010;678:119-123 Crossref
  • [63] J.S. Myers, J. Pierce, T. Pazdernik. Neurotoxicology of chemotherapy in relation to cytokine release, the blood-brain barrier, and cognitive impairment. Oncol Nurs Forum. 2008;35:916-920 Crossref
  • [64] J. Licinio, M.A. Kling, P. Hauser. Cytokines and brain function: relevance to interferon-alpha-induced mood and cognitive changes. Semin Oncol. 1998;25:30-38
  • [65] S. Salim, G. Chugh, M. Asghar. Inflammation in anxiety. Adv Protein Chem Struct Biol. 2012;88:1-25 Crossref
  • [66] G. Patki, N. Solanki, F. Atrooz, F. Allam, S. Salim. Depression, anxiety-like behavior and memory impairment are associated with increased oxidative stress and inflammation in a rat model of social stress. Brain Res. 2013;1539:73-86 Crossref
  • [67] A. Liberati, D.G. Altman, J. Tetzlaff, C. Mulrow, P.C. Gøtzsche, J.P. Ioannidis, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ. 2009;339:b2700 Crossref


a Department of Pharmacy, National University of Singapore, Singapore

b Human Genetics, Genome Institute of Singapore, Singapore

c Department of Pharmacy, National Cancer Centre Singapore, Singapore

lowast Corresponding author at: Department of Pharmacy, National University of Singapore, Block S4A, Level 3, 18 Science Drive 4, Singapore 117543, Singapore. Tel.: +65 6516 7814; fax: +65 6779 1554.

Search this site

Stay up-to-date with our monthly e-alert

If you want to regularly receive information on what is happening in Quality of Life in Oncology research sign up to our e-alert.

Subscribe »

QOL (Quality of Life) newsletter e-alert

NEW! Free access to the digital version of a new publication in Cancer Supportive Care

Cancer cachexia: mechanisms and progress in treatment

Authors: Egidio Del Fabbro, Kenneth Fearon, Florian Strasser

This book was supported by an educational grant from Helsinn Healthcare SA.

Featured videos

Quality of Life promotional video

Made possible by an educational grant from Helsinn

Helsinn does not have any influence on the content and all items are subject to independent peer and editorial review

Society Partners

European Cancer Organisation Logo