dove medical press web viewbmi, smoking status, alcohol consumption, ethnicity, rheumatoid...
TRANSCRIPT
Bisphosphonates and primary breast cancer risk: an updated systematic review and meta-
analysis involving 963,995 women
Yupeng Liu, PhD; Xiaosan Zhang, PhD; Hongru Sun, MSc; Shu Zhao, PhD; Yuxue Zhang, MSc; Dapeng Li, MSc;
Qingyuan Zhang, PhD; Yashuang Zhao, PhD
SUPPLEMENTARY MATERIAL
Correspondence to: Dr. Yupeng Liu Department of Preventive Medicine, School of Public Health and Management
Wenzhou Medical UniversityWenzhou 325035, P. R. China
Tel/Fax: +86 (0) 577 8669 9182E-mail: [email protected] Professor Yashuang Zhao
Department of Epidemiology, School of Public HealthHarbin Medical UniversityHarbin 150086, P. R. China
Tel/Fax: +86 (0) 451 8750 2823E-mail: [email protected]
or Professor Qingyuan Zhang Department of Medical Oncology, The Third Affiliated Hospital
Harbin Medical UniversityHarbin150081, P. R. China
Tel/Fax: +86 (0) 451 8629 8116E-mail: [email protected]
Table of Contents Pages
Supplementary Discussion 1
Supplementary Tables 2-10
1 eTable 1. Characteristics of cohort studies included in this meta-analysis 2
2 eTable 2. Characteristics of case-control studies included in this meta-analysis 3
3 eTable 3. Characteristics of randomized controlled trials addressing the association between bisphosphonates and the risk of primary breast cancer
4
4 eTable 4. Sensitivity analyses by comparisons of the pooled results of the maximally and minimally adjusted effect estimates of the association between bisphosphonates and breast cancer risk
5
5 eTable 5. Sensitivity analyses excluding the study that has the most relative weight in subgroups 6
6 eTable 6. Characteristics of cohort studies addressing the association between bisphosphonates and the risk of contralateral breast cancer
7
7 eTable 7. Characteristics of case-control studies addressing the association between bisphosphonates and the risk of contralateral breast cancer
8
8 eTable 8. Publication bias 9
9 eTable 9. Levels of evidence for associations between bisphosphonates and primary breast cancer risk 10
Supplementary Figures 11-22
10 eFigure 1. Forest plot for the association between bisphosphonates and the risk of primary invasive breast cancer
11
11 eFigure 2. Forest plot for the association between bisphosphonates and primary breast cancer risk in postmenopausal women
12
12 eFigure 3. Subgroup analyses of the association between bisphosphonates and primary breast cancer risk according to duration of bisphosphonate use.
13
13 eFigure 4. Cumulative meta-analysis of the association between bisphosphonates and primary breast cancer risk according to duration of bisphosphonate use
14
14 eFigure 5. A sensitivity analysis by excluding retrospective case-control studies 15
15 eFigure 6. A sensitivity analysis by excluding randomized controlled trials 16
16 eFigure 7. Forest plot for the association between bisphosphonates and the risk of contralateral breast cancer in women with primary breast cancer
17
17 eFigure 8. A sensitivity analysis by combining the results from studies addressing the associations of bisphosphonates with the risk of both primary and contralateral breast cancer
18
18 eFigure 9. Meta-regression analyses of the association between bisphosphonates and primary breast cancer risk
19-20
19 eFigure 10. Funnel plots of potential publication 21
20 eFigure 11. Risk of bias of randomized controlled trials 22
Appendix 23-25
21 Appendix A. PRISMA Checklist 23-24
22 Appendix B. Methodological quality score of the cohort and case-control studies 25
Supplementary Discussion
Additionally, we assessed the statistical power of these two RCTs by using the software of PASS version 11.0.7 (NCSS
LLC., USA). In the Hue’s study, the power of statistical analyses for the association between bisphosphonates and BCa
was only 19.53% and 8.58% for FIT and HORIZON-PFT, respectively, based on the sample size (6194 participants in
FIT and 7580 in HORIZON-PFT), the incidence of breast cancer in the control groups (1.49% in FIT and 0.77% in
HORIZON-PFT), and the effect estimates (1.24 for FIT and 1.15 for HORIZON-PFT). Even though pooling the data
from the two trials, the power was only 21.71%, indicating an overall 78.29% risk of type II error. Given the limited
number of BCa patients during the follow-up period (165 patients from the combined dataset), this low power is not
surprising. In contrast, the included observational studies, especially prospective cohort studies, with a large sample size,
have an extremely high power of statistical analyses to detect the association. The pooling analysis of cohort studies had
a power of 99.81% (with a two-sided alpha of 0.05) to detect a 10% reduction of the risk of developing primary BCa in
the bisphosphonate users.
1 / 25
SUPPLEMENTARY TABLES
eTable 1. Characteristics of cohort studies included in this meta-analysisStudy, Country
Population characteristicsmean age (SD)/median age (range), years
Study period,Follow-up duration mean (SD) , years
No. of BPs-users(No. of cases)
No. of nonusers(No. of cases)
Types of BPs, No. of users (%)
Data sources of BPs prescription
BCa diagnosis Adjusted/controlled factors Quality Score* (%)
Chiang 2012 16
Taiwan, ChinaPostmenopausal women aged over 55BPs-users: 73.5(8.4)Non-users: 73.4(8.4)
1998-2010BPs-users: 4.3(2.5)Non-users: 4.9(2.6)
6906 (44) 20697 (165) All alendronate,Alendronate: 6909 (100.0)
NHIRD Hospital diagnosis,mostly confirmed by the pathological examination
Age, osteoporosis, hypertension, diabetes, benign breast disease, morbid obesity, statin use, chronic obstructive pulmonary disease, oestrogen use, dyslipidemia, chronic kidney disease, coronaryartery disease, colorectal polyp, etc.
80
Cardwell 2011 17
the United KingdomUKGPRD81% female with a mean age of 70 Both groups: 70.0(11.4)
1996-2006BPs-users: 4.5 (2.6) Non-users: 4.4 (2.6)
34049 (369) 34049 (501) Any BPs,Alendronate: 23834 (70.0)Nitrogen containingBPs: 29623 (87.0)
UKGPRD Hospital diagnosis,identified from the UKGPRD for diagnosis
Age, sex, BMI, alcohol, smoking, NSAID prescription, glucocorticoid steroid, vitamin D prescription and calcium, history of osteoporosis or osteopenia etc.
86
Vestergaard 2011 24
DenmarkWomen aged over 40Both groups: 71.1(10.7)
1996-2006NS
81329 (884) 261322 (3465) Alendronate: 45162(51.8)Etidronate: 34103 (39.2)Risedronate: 1156 (1.3)Ibandronate: 540 (0.6)Clodronate: 339 (0.4)Pamidronate: 21 (<0.1)Zoledronate: 8 (<0.1)
DNPD Hospital diagnosis,identified from the Danish cancer registry
Age, use of systemic hormone therapy, irradiation, chemotherapy, alcoholism, etc.
78
Chlebowski 2010 25
the United StatesPostmenopausal womenBoth groups: 63.8 (50-79)
1993-2005Both groups: 8.5
2816 (64) 151952 (5092) Alendronate: 2527 (89.7)Etidronate: 285 (10.1)Pamidronate: 1 (<0.1)Tiludronate: 1 (<0.1)More than one: 2 (0.1)
An interview-administered questionnaire
Hospital diagnosis, verified at each clinic by medical record and pathology report
Age, ethnicity, smoking, alcohol use, physical activity, BMI, mammogram in the last 2 years, chemotherapy, calcium, vitamin D, 5-year risk of hip fracture, Gail 5-year risk of BCa, etc.
81
Fournier 201718
FrenchThe E3N cohort
Postmenopausal Women 62.8 (6.4)BPs-users: 64.8 (6.6)Non-users: 62.3 (6.3)
2004-2011Both groups: 7.2 (1.7)
12935(308) 51503(2099) Any BPs†, Alendronate: 14071Risedronate: 11097Ibandronate: 4246Etidronate: 1461Zoledronate: 693Tiludronate: 2
A series of questionnaires
Self-report using a follow-up questionnaire and 95% cases with pathology reports
BMI, years of schooling, alcohol intake, personal history of fractures, osteoporosis and bone densitometries, parity and age at first birth, use of oral contraceptives, time since menopause, history of breast cancer in first-degree relatives,personal history of benign breast disease, hormone replacement therapy use, self-report of a mammogram performed during the previous follow-up cycle, raloxifene, and vitamin D
89
Abbreviations: BCa, breast cancer; BMI, body mass index; BPs, bisphosphonates; CIPD, the Catastrophic Illness Patient Database; DNPD, the Danish National Prescription Database; NHIRD, the National Health Insurance Research Database; NS, not stated; NSAID, nonsteroidal anti-inflammatory drug; SD, standard deviation; UKGPRD, the United Kingdom General Practice Research Database; The E3N cohort, the French prospective Etude Epidemiologique aupres de femmes de la Mutuelle Generale de l'Education Nationale cohort.*The quality score of each study is presented as a percentage of the maximum score.†Did not report the numbers of BPs users but reported the number of person-years of BPs users.
2 / 25
eTable 2. Characteristics of case-control studies included in this meta-analysis
Study, Country,Study design
Population characteristicsmean age (SD)/median age (range), years
Study period No. of cases(No. of BPs-users)
No.of controls(No. of BPs-users)
Types of BPs, No. of users (%)
Data sources of BPs prescription
BCa diagnosis Adjusted/controlled factors Quality Score* (%)
Rennert 2010 34
IsraelCase-control(Population-based individually matched by age, residence, and ethnic group)
Postmenopausal womenCase: 63.6Control: 65.6
2000- 1832 (193) 2207 (326) Alendronate: 519 (86.7)
CHS Hospital diagnosis,identified independently by the diagnosing hospitals
Age, residence, ethnicity, BMI, age at first pregnancy, family history of BCa, first degree, sports activity, fruits consumption, etc.
78
Newcomb 2010 33
the United StatesCase-control(Population-based, frequency matched by age)
Women aged 20–69Case: 54.2(8.9)Control: 54.8(8.9)
2003-2006 2936 (106) 2975 (161) Any BPs† An interview-administered questionnaire
Hospital diagnosis,identified from Wisconsin'smandatory cancer registry
Age at diagnosis, parity, age at first live birth, family history of breast cancer, BMI, menopausal status, type of hormone use, mammography, osteoporosis, smoking, height change etc.
73
Vinogradova 201319
the United Kingdoma nested case-control study from the QResearch and CPRD databases(Population-based individually matched by age, sex, practice and calendar year)
QResearch:Case: 69.4(9.8)Control: 69.4(9.7)CPRD:Case: 69.9(10.4)Control: 69.7(10.2)
1997-2011 QResearch:24489 (3827)CPRD:25444 (3769)
QResearch:113945 (17883)CPRD:118835 (17490)
Any BPsAlendronate Etidronate Risedronate Ibandronate
BNF From the two largest primary-care databases in the UK: QResearch and CPRD
BMI, smoking status, alcohol consumption, ethnicity, rheumatoid arthritis, osteoporosis and fractures, use of other osteoporosis drugs, vitamin D, NSAIDs, corticosteroids, acid-lowering drugs, years of data, family history of breast cancer, benign breast disease, use of oral contraceptives and hormone replacement therapy.
75
Abbreviations: BCa, breast cancer; BMI, body mass index; BNF, British National Formulary; BPs, bisphosphonates; CHS, the Clalit Health Services pharmacy records; NSAID, nonsteroidal anti-inflammatory drugs; SD, standard deviation; *The quality score of each study is presented as a percentage of the maximum score.†Did not report the types and numbers of BPs users.
3 / 25
eTable 3. Characteristics of randomized controlled trials addressing the association between bisphosphonates and the risk of primary breast cancer
Study, Country
Population characteristicsmean age (SD)/median age (range), years
Study period,Follow-up duration mean (SD) , years
No. of BPs-users(No. of Events)
No. of nonusers(No. of Events) BPs treatment BCa diagnosis Outcomes
/End-pointsJadad Score
Hue 2014 14
FIT a randomized double-blind placebo-controlled interventional trialthe United States
Postmenopausal women 55-81Alendronate: 68.1(6.2)Placebo: 68.2(6.1)
1992/1993-3.8 (0.8)
3103 (57) 3091 (46) Daily oral alendronate sodium (5mg/day for 2 years and 10 mg/day for 2 years)
Adverse events reports and supporting medical records†
Fracture 4
Hue 2014 14
HORIZON-PFT the United States
Postmenopausal women 65-89Zoledronic acid: 73.1 (5.3)Placebo: 73.0 (5.4)
2002-20112.8 (0.6)
3792 (33) 3788 (29) Annual intravenous infusions of zoledronic acid 5mg for 3 years
Adverse events reports and supporting medical records†
Fracture and BMD 4
†FIT and HORIZON-PFT were not initially designed to study breast cancer outcomes and identified new breast cancer cases using adverse events reports and supporting medical records.
4 / 25
eTable 4. Sensitivity analyses by comparisons of the pooled results of the maximally and minimally adjusted effect estimates of the association between bisphosphonates and breast cancer risk* Subgroups No. of
studies Effect estimate (95% CI) I2 (%) P-value for heterogeneity
Confounding RR† (P-value)
Total
Maximal 10 0.879 (0.826-0.936) 35.53 0.1241.060, (0.256)
Minimal 10 0.829 (0.766-0.898) 66.16 0.002
Case-control study
Maximal 3 0.834 (0.711-0.979) 61.79 0.073 1.090, (0.511)
Minimal 3 0.765 (0.625-0.936) 80.52 0.006
Cohort study
Maximal 5 0.874 (0.803-0.951) 33.18 0.2001.059, (0.373)
Minimal 5 0.825 (0.750-0.906) 53.89 0.070
Randomized controlled trials
Maximal 2 1.127 (0.808-1.573) 0.00 0.918 0.935, (0.772)
Minimal 2 1.205 (0.887-1.637) 0.00 0.815
Invasive BCa
Maximal 6 0.829 (0.692-0.994) 59.54 0.030 1.067, (0.602)
Minimal 6 0.770 (0.624-0.951) 75.53 0.001
Postmenopausal
Maximal 7 0.888 (0.798-0.989) 30.50 0.1951.043, (0.642)
Minimal 7 0.851 (0.737-0.982) 67.47 0.005
Abbreviations: BCa, breast cancer; BPs, bisphosphonates; CI, confidence interval.*We used a random-effects model to pool the effect estimates and therefore be more conservative.†Confounding RR was defined as the ratio of the pooled results of the maximally and the minimally adjusted effect estimates. If the P-value for confounding RR was less than 0.05, the impact of adjustment for confounders on the pooled results was statistically significant; otherwise, the confounding effect of the adjusted factors was small or null. For protective factors, if the confounding RR <1, it indicates that the adjusted confounders have obscured an inverse association; if the confounding RR >1, it indicates that the adjusted confounders have exaggerated an inverse association.
5 / 25
eTable 5. Sensitivity analyses excluding the study that has the most relative weight in subgroups*
Subgroups No. of studies Effect estimate (95% CI)Heterogeneity between-study
P-value for heterogeneity
between subgroupsI2 (%) P-value
Total 10 0.879 (0.826-0.936) 35.53 0.1240.738
Total† 9 0.864 (0.798-0.935) 30.08 0.178
Cohort study
Subtotal 5 0.874 (0.803-0.951) 33.18 0.2000.912
Subtotal† 4 0.882 (0.769-1.011) 47.30 0.128
Invasive BCa
Subtotal 6 0.829 (0.692-0.994) 59.54 0.030 0.647
Subtotal† 5 0.780 (0.646-0.941) 41.31 0.146
Postmenopausal
Subtotal 7 0.888 (0.798-0.989) 30.50 0.1950.539
Subtotal† 6 0.846 (0.757-0.946) 9.91 0.353
Duration of BPs use (year)
< 1
Subtotal 5 0.904 (0.842-0.970) 0.00 0.522 0.724
Subtotal† 4 0.873 (0.729-1.046) 1.33 0.385
≥1 (including 1-2, 2-3, 3-4, and >4)
Subtotal 19 0.745 (0.658-0.844) 74.61 <0.0010.779
Subtotal† 18 0.727 (0.647-0.818) 57.40 0.001
1-2
Subtotal 5 0.690 (0.534-0.893) 87.00 <0.001 0.561
Subtotal† 4 0.629 (0.527-0.750) 42.47 0.157
2-3
Subtotal 4 0.732 (0.554-0.967) 72.82 0.012 0.694
Subtotal† 3 0.797 (0.579-1.098) 67.65 0.045
3-4
Subtotal 6 0.862 (0.682-1.090) 59.40 0.0310.555
Subtotal† 5 0.941 (0.792-1.119) 0.00 0.665
>4
Subtotal 4 0.678 (0.525-0.875) 38.45 0.181 0.791
Subtotal† 3 0.721 (0.494-1.051) 54.51 0.111
Abbreviations: BCa, breast cancer; BPs, bisphosphonates; CI, confidence interval.*This sensitivity analysis was conducted only when the number of included studies in subgroup analyses was not less than 4. We used a random-effects model to pool the effect estimates and therefore be more conservative.†The pooled results of sensitivity analyses excluding the study that has the most relative weight in subgroups.
6 / 25
eTable 6. Characteristics of cohort studies addressing the association between bisphosphonates and the risk of contralateral breast cancerStudy, Country
Population characteristicsmean age (SD)/median age (range), years
Study period,Follow-up duration mean (SD) , years
No. of BPs-users(No. of cases)
No. of nonusers(No. of cases)
Types of BPs, No. of users (%)
Data sources of BPs prescription
BCa diagnosis Adjusted/controlled factors Quality Score* (%)
Kwan 201636
the United StatesThe KPNC and KPSC cohorts
KPNC 59.6 (12.5)BPs-users: 64.7 (11.0)Non-users: 58.8 (12.5)KPSC 59.7 (12.3)BPs-users: 64.5 (10.7)Non-users: 58.1 (12.3)
1996-2009KPNC BPs-users: 3.59Non-users: 6.67KPSC BPs-users: 3.41Non-users: 7.06
KPNC 1282 (35)KPSC1973 (26)
KPNC 7575 (199)KPSC5951 (172)
Any BPs,Alendronate: most commonly prescribed (>93 %)
From the Kaiser Permanente insurance records
Hospital diagnosis,from thecancer registries of the United States
Age at diagnosis, year of diagnosis, race/ethnicity, Charlson comorbidity index, AJCC stage, primary therapy, chemotherapy, ER/PR status, Her2 status, post-BC diagnosis osteopenia, osteoporosis, or fracture, and tamoxifen and aromatase inhibitor use
87
Korde 201837
the United StatesThe Quilt cohort joining the cases recruited into threeprevious population-based studies of breast cancer, the CARE, PACE and EMF studies
64.2 (45-79) 1993-201511.8 (2.5-17.6)
302 (23) 1511 (139) Any BPs,Alendronate 272(90.0)Risendronate 40(13.0)Zoledronate 12(4.0)
From medical record reviews
From the Cancer Surveillance System, the population-based SEER registry
Age at diagnosis and source study (the original three studies through which women were ascertained)
82
Abbreviations: BCa, breast cancer; BPs, bisphosphonates; KPNC, the Kaiser Permanente Northern California cohort; KPSC, the Kaiser Permanente Southern California cohort; SD, standard deviation. *The quality score of each study is presented as a percentage of the maximum score.
7 / 25
eTable 7. Characteristics of case-control studies addressing the association between bisphosphonates and the risk of contralateral breast cancer
Study, Country,Study design
Population characteristicsmean age (SD)/median age (range), years
Study period No. of cases(No. of BPs-users)
No.of controls(No. of BPs-users)
Types of BPs, No. of users (%)
Data sources of BPs prescription
BCa diagnosis Adjusted/controlled factors Quality Score* (%)
Monsees 201135
the United Statesa nested case-control study from the SEER database(Population-based individually matched by age atand year of first breast cancer diagnosis, ethnicity, historic stage of first breast cancer )
primary invasive, localized or regional SEER historic stage ER+ breast cancer women61.1 (40-79)
1990-2007 351 (32) 662 (85) Any nitrogenous BPsAlendronate:106(88.0)
The medical records of the oncology and/or primary care provider
SEER age and year of diagnosis with first primary breast cancer,SEER historic stage of the first primary breast cancer, county of residence, race/ethnicity and BMI at first breast cancer diagnosis
84
Abbreviations: BCa, breast cancer; BMI, body mass index; BPs, bisphosphonates; SD, standard deviation; SEER, the Surveillance Epidemiology and End Results.*The quality score of each study is presented as a percentage of the maximum score.
8 / 25
eTable 8. Publication bias
Subgroups No. of studies
P-value Summarized effect estimates (95% CI)*
For Begg's rank
correlation test
For Egger's linear
regression test
Adjusted
with trim and fillUnadjusted
Total 10 0.721 0.698 0.876 (0.822-0.932) 0.879 (0.826-0.936)
Case-Control study 3 0.296 0.117 0.920 (0.876-0.967) 0.834 (0.711-0.979)
Cohort study 5 0.807 0.881 0.839 (0.767-0.917) 0.877 (0.811-0.949)
Randomized controlled trial 2 NA NA NA 1.127 (0.808-1.573)
Invasive BCa 6 0.851 0.966 0.766 (0.637-0.922) 0.829 (0.692-0.994)
Postmenopausal 7 0.230 0.533 0.866 (0.779-0.962) 0.888 (0.798-0.989)
Duration of BPs treatment (year)
<1 5 0.624 0.835 0.904 (0.842-0.970) 0.904 (0.842-0.970)
Subtotal ≥1 19 0.132 0.053 0.734 (0.648-0.831) 0.745 (0.658-0.844)
1-2 5 0.462 0.061 0.818 (0.659-1.014) 0.690 (0.534-0.893)
2-3 4 0.734 0.915 0.796 (0.608-1.041) 0.732 (0.554-0.967)
3-4 6 0.060 0.014 0.826 (0.665-1.026) 0.862 (0.682-1.090)
>4 4 0.308 0.324 0.628 (0.484-0.816) 0.678 (0.525-0.875)
Abbreviations: BCa, breast cancer; BPs, bisphosphonates; CI, confidence interval; NA, not applicable. *The summarized effect estimate was calculated with a random-effects model.
9 / 25
eTable 9. Levels of evidence for associations between bisphosphonates and primary breast cancer risk
SubgroupsNo. of studies (%)
SubtotalLevel
of evidenceDecreased risk Increased risk No association
Total 5 (50) 0 5 (50) 10 Indecisive
Case-Control study 2 (67) 0 1(33) 3 Moderate
Cohort study 3 (60) 0 2 (40) 5 Moderate
RCTs 0 0 2(100) 2 Indecisive
Invasive BCa 3 (50) 0 3(50) 6 Indecisive
Postmenopausal 3 (43) 0 4 (57) 7 Indecisive
Duration of bisphosphonnate use (year)
< 1 2 (40) 0 3 (60) 5 Nil
≥ 1 (subtotal) 13 (68) 0 6 (32) 19 Strong
Abbreviations: BCa, breast cancer; BPs, bisphosphonates.
10 / 25
SUPPLEMENTARY FIGURES
eFigure 1. Forest plot for the association between bisphosphonates and the risk of primary
invasive breast cancer. Square markers indicate effect sizes of each study; horizontal lines, the
95% confidence intervals. The diamond data marker indicates the summarized effect size. The
vertical solid line indicates the overall pooled effect. Note that the studies are ranked in order of
their relative weights from random effects analysis. RR, relative risk.
Study
Fournier et al,18 2017
Overall (I2 = 59.54%, P = 0.03)
Chlebowski et al,25 2010
Rennert et al,34 2010
Newcomb et al,33 2010
RR (95% CI)
0.97 (0.83, 1.14)
0.83 (0.69, 0.99)
0.68 (0.52, 0.89)
0.70 (0.54, 0.90)
0.70 (0.54, 0.91)
Weight %
24.59
18.18
19.11
18.31
1.5 1 2Test for overall effect: z = -2.03; P = 0.04
Favor bisphosphonatesusers
Favor non-users
RR (95% CI)
100
Hue et al,14 2014_HORIZON-PFT
Hue et al,14 2014_FIT
1.15 (0.70, 1.90)
1.11 (0.71, 1.73)
11 / 25
eFigure 2. Forest plot for the association between bisphosphonates and primary breast cancer
risk in postmenopausal women. Square markers indicate effect sizes of each study; horizontal
lines, the 95% confidence intervals. The diamond data marker indicates the summarized effect size.
The vertical solid line indicates the overall pooled effect. Note that the studies are ranked in order of
their relative weights from random effects analysis. RR, relative risk.
Overall (I2 = 30.50%, P = 0.20)
Cardwell et al,17 2011
Chlebowski et al,25 2010
Rennert et al,34 2010
Fournier et al,18 2017
Study
Chiang et al,16 2012
0.89 (0.80, 0.99)
0.75 (0.63, 0.89)
0.90 (0.72, 1.12)
RR (95% CI)
0.80 (0.65, 0.99)
0.98 (0.85, 1.12)
0.92 (0.66, 1.28)
21.68
15.97
Weight %
17.05
27.25
8.65
1.5 1 2
Favor bisphosphonatesusers
Favor non-users
RR (95% CI)
Test for overall effect: z = -2.17; P = 0.03
100
Hue et al,14 2014 -FIT
Hue et al,14 2014 -HORIZON-PFT
1.11 (0.71, 1.73)
1.15 (0.70, 1.90)
12 / 25
eFigure 3. Subgroup analyses of the association between bisphosphonates and primary breast cancer
risk according to duration of bisphosphonate use. <1 indicates the subgroup of bisphosphonate users with
less than one year use of bisphosphonates; 1-2, one year or more but less than two years; 2-3, two years or
more but less than three years; 3-4, three years or more but less than four years; ≥4, 4 years or more. Square
markers indicate effect sizes of each study; horizontal lines, the 95% confidence intervals. The diamond data
marker indicates the summarized effect size. Note that the studies are ranked in order of their relative
weights from random effects analysis. RR, relative risk.
13 / 25
Cardwell et al,17 2012
Rennert et al,34 2010
Study or Subgroup
Cardwell et al,17 2012
Fournier et al,18 2017
Subtotal_<1 yr (I2 = 0.00%, P = 0.52)
Rennert et al,34 2010
Rennert et al,34 2010
Fournier et al,18 2017
Subtotal_>4 yr (I2 = 38.45%, P = 0.18)
Rennert et al,34 2010
Vinogradova et al,19 2013
Cardwell et al,17 2012
Rennert et al,34 2010
Newcomb et al,33 2010
Subtotal_3-4 yr (I2 = 59.40%, P = 0.03)
Fournier et al,18 2017
Subtotal_2-3 yr (I2 = 72.82%, P = 0.01)
Vinogradova et al,19 2013
Rennert et al,34 2010
Newcomb et al,33 2010
Cardwell et al,17 2012Chlebowski et al,25 2010
Fournier et al,18 2017
Chlebowski et al,25 2010
Subtotal_1-2 yr (I2 = 87.88%, P < 0.01)
Chlebowski et al,25 2010
Newcomb et al,33 2010
0.81 (0.57, 1.16)
0.57 (0.44, 0.74)
0.74 (0.61, 0.90)
1.04 (0.68, 1.59)
0.90 (0.84, 0.97)
0.61 (0.50, 0.75)
0.59 (0.47, 0.75)
0.77 (0.57, 1.04)
0.68 (0.52, 0.88)
RR (95% CI)
0.51 (0.34, 0.76)
0.93 (0.87, 0.99)
0.71 (0.54, 0.94)
0.93 (0.61, 1.42)
0.69 (0.41, 1.17)
0.86 (0.68, 1.09)
1.07 (0.81, 1.41)
0.73 (0.55, 0.97)
0.91 (0.84, 0.98)
0.61 (0.45, 0.83)
0.63 (0.42, 0.95)
0.94 (0.57, 1.55)0.83 (0.54, 1.28)
1.14 (0.78, 1.66)
0.50 (0.38, 0.66)
0.69 (0.53, 0.89)
0.86 (0.64, 1.16)
0.78 (0.54, 1.12)
17.42
21.38
21.89
14.87
21.65
28.07
5.54
Weight %
25.51
24.83
25.98
2.81
12.32
25.98
84.37
33.12
19.97
18.8722.50
3.51
19.31
19.60
3.77
1.25 .5 1 2
Test for subtotal effect:z = -2.79; P = 0.005
Test for subtotal effect:z = -2.82; P = 0.005
Test for subtotal effect:z = -2.19; P = 0.03
Test for subtotal effect:z = -1.24; P = 0.22
Test for subtotal effect:z = -2.98; P = 0.003
Test for subgroup differences:P = 0.007
RR (95% CI)
Favor bisphosphonatesusers
Favor non-users
100
100
100
100
100
Hue et al,
14 / 25
eFigure 4. Cumulative meta-analysis of the association between bisphosphonates and primary
breast cancer risk according to duration of bisphosphonate use. Square markers indicate effect
sizes of each study; horizontal lines, the 95% confidence intervals. The vertical solid line indicates
the overall pooled effect. RR, relative risk.
Rennert,34 2010_>5yr
Newcomb,33 2010_>2yr
Cardwell,17 2011_>4yr
Fournier,18 2017_<0.5yr
Rennert,34 2010_4-5yrFournier,18 2017_≥3yr
Vinogradova,19 2013_≥1yr
Fournier,18 2017_0.5-1yr
Rennert,34 2010_1-2yr
Rennert,34 2010_<1yr
Chlebowski,25 2010_<2yr
Rennert,34 2010_2-3yr
Chlebowski,25 2010_2-5yr
Newcomb,33 2010_<1yr
Cardwell,17 2011_2-3yr
Fournier,18 2017_1-3yr
Chlebowski,25 2010_>5yr
Newcomb,33 2010_1-2yr
Cardwell,17 2011_3-4yrRennert,34 2010_3-4yr
Cardwell,17 2011_1-2yr
Vinogradova,19 2013_<1yr
Study
0.77 (0.70, 0.85)
0.77 (0.69, 0.86)
0.78 (0.71, 0.86)
0.77 (0.57, 1.04)
RR (95% CI)
0.78 (0.71, 0.86)0.79 (0.71, 0.87)
0.79 (0.69, 0.89)
0.92 (0.63, 1.35)
0.81 (0.69, 0.96)
0.91 (0.72, 1.15)
0.76 (0.64, 0.90)
0.78 (0.69, 0.88)
0.78 (0.70, 0.86)
0.87 (0.73, 1.05)
0.78 (0.69, 0.87)
0.81 (0.72, 0.91)
0.78 (0.71, 0.86)
0.90 (0.84, 0.97)
0.76 (0.69, 0.85)0.76 (0.68, 0.85)
0.80 (0.69, 0.93)
0.90 (0.84, 0.97)
100
65.71
93.76
4.37
Weight %
91.2987.06
47.87
7.85
29.93
10.91
40.32
57.84
70.07
14.52
62.52
52.55
96.70
24.26
78.7074.98
35.73
21.96
1.5 1 2
Favor bisphosphonatesusers
Favor non-users
RR (95% CI)
Hue,14 2014_HORIZON-PFT_3-4yrHue,14 2014_FIT_3-4yr
0.78 (0.70, 0.86)
15 / 25
eFigure 5. A sensitivity analysis by excluding retrospective case-control studies. Square
markers indicate effect sizes of each study; horizontal lines, the 95% confidence intervals. The
diamond data marker indicates the summarized effect size. The vertical solid line indicates the
overall pooled effect. Note that the studies are ranked in order of their relative weights from random
effects analysis. HORIZON-PFT, Health Outcomes and Reduced Incidence with Zoledronic Acid
Once Yearly–Pivotal Fracture Trial; FIT, Fracture Intervention Trial; RR, relative risk.
Overall (I2 = 26.75%, P = 0.215)
Study
Fournier et al,18 2017
Chlebowski et al,25 2010
Hue et al,14 2014_FIT
Chiang et al,16 2012
Hue et al,14 2014_HORIZON-PFT
Vestergaard et al,24 2011
Cardwell et al,17 2011
0.90 (0.85, 0.95)
0.98 (0.85, 1.12)
0.90 (0.72, 1.12)
1.11 (0.71, 1.73)
0.92 (0.66, 1.28)
1.15 (0.70, 1.90)
RR (95% CI)
0.86 (0.80, 0.93)
0.75 (0.63, 0.89)
13.43
6.17
1.66
2.93
1.32
Weight %
28.39
9.41
1.5 1 2
100
Favor bisphosphonatesusers
Favor non-users
Test for overall effect: z = -3.68; P < 0.001
RR (95% CI)
Vinogradova et al, 19 2013 0.92 (0.87, 0.97)
16 / 25
eFigure 6. A sensitivity analysis by excluding randomized controlled trials. Square markers
indicate effect sizes of each study; horizontal lines, the 95% confidence intervals. The diamond data
marker indicates the summarized effect size. The vertical solid line indicates the overall pooled
effect. Note that the studies are ranked in order of their relative weights from random effects
analysis. RR, relative risk.
Overall (I2 = 41.77%, P = 0.10)
Newcomb et al, 33 2010
Vestergaard et al, 24 2011
Chiang et al, 16 2012
Fournier et al, 18 2017
Cardwell et al, 17 2011
Rennert et al, 34 2010
Vinogradova et al, 19 2013
Chlebowski et al, 25 2010
Subtotal_Case-control (I2 = 61.79%, P = 0.07)
Subtotal_Cohort (I2 = 33.18%, P = 0.20)
Study
0.87 (0.82, 0.93)
0.70 (0.54, 0.91)
0.86 (0.80, 0.93)
RR (95% CI)
0.92 (0.66, 1.28)
0.98 (0.85, 1.12)
0.75 (0.63, 0.89)
0.80 (0.65, 0.99)
0.92 (0.87, 0.97)
0.90 (0.72, 1.12)
0.83 (0.71, 0.98)
0.87 (0.80, 0.95)
5.05
24.39
Weight %
3.43
13.67
10.09
7.47
28.99
6.91
1.5 1 2
Favor bisphosphonatesusers
Favor non-users
Test for subgroup differences: P = 0.37
Test for overall effect: z = -4.18;
17 / 25
eFigure 7. Forest plot for the association between bisphosphonates and the risk of
contralateral breast cancer in women with primary breast cancer. Square markers indicate
effect sizes of each study; horizontal lines, the 95% confidence intervals. The diamond data marker
indicates the summarized effect size of the observational studies included in this meta-analysis and
the vertical solid line indicates the overall pooled effect. The studies are ranked in order of their
relative weights from random effects analysis. Note that the EBCTCG’ study (Early Breast Cancer
Trialists’ Collaborative Group) is a meta-analysis of individual patient data from randomized
controlled trials of adjuvant bisphosphonate treatment in early breast cancer patients. RR, relative
risk.
Korde et al,37 2018
Monsees et al,35 2011
Overall (I2 = 0.00%, P = 0.63)
Study
Kwan et al,36 2016
RR (95% CI)
0.75 (0.47, 1.19)
0.77 (0.49, 1.19)
0.84 (0.66, 1.07)
0.96 (0.67, 1.38)
Weight %
26.87
29.52
43.61
1.5 2
Favor bisphosphonatesusers
Favor non-users
Test for overall effect: z = -1.43; P = 0.15
EBCTCG,43 2015 0.96 (0.74, 1.25)
18 / 25
eFigure 8. A sensitivity analysis by combining the results from studies addressing the
associations of bisphosphonates with the risk of both primary and contralateral breast cancer.
Square markers indicate effect sizes of each study; horizontal lines, the 95% confidence intervals.
The diamond data marker indicates the summarized effect size. The vertical solid line indicates the
overall pooled effect. Note that the studies are ranked in order of their relative weights from random
effects analysis EBCTCG, Early Breast Cancer Trialists’ Collaborative Group; HORIZON-PFT,
Health Outcomes and Reduced Incidence with Zoledronic Acid Once Yearly–Pivotal Fracture Trial;
FIT, Fracture Intervention Trial; RR, relative risk.
Rennert et al, 34 2010
Hue et al, 14 2014_HORIZON-PFT
Study
Hue et al, 14 2014_FIT
Chiang et al, 16 2012
Newcomb et al, 33 2010
Kwan et al, 36 2016
Fournier et al, 18 2017
EBCTCG, 43 2015
Chlebowski et al, 25 2010
Korde et al, 37 2018
Cardwell et al, 17 2011
Overall (I2 = 15.74%, P = 0.28)
Vestergaard et al, 24 2011
Vinogradova et al, 19 2013
Monsees et al, 35 2011
0.80 (0.65, 0.99)
1.15 (0.70, 1.90)
RR (95% CI)
1.11 (0.71, 1.73)
0.92 (0.66, 1.28)
0.70 (0.54, 0.91)
0.96 (0.67, 1.38)
0.98 (0.85, 1.12)
0.96 (0.74, 1.25)
0.90 (0.72, 1.12)
0.75 (0.47, 1.19)
0.75 (0.63, 0.89)
0.88 (0.84, 0.93)
0.86 (0.80, 0.93)
0.92 (0.87, 0.97)
0.77 (0.49, 1.19)
5.18
0.99
Weight %
1.25
2.22
3.36
1.88
10.65
3.46
4.74
1.18
7.33
24.09
32.38
1.29
1.5 1 2
100
Test for overall effect: z = -4.08;
19 / 25
eFigure 9. Meta-regression analyses of the association between bisphosphonates and primary
breast cancer risk according to (A) the mean age of participants, (B) the quality score, (C) the
total number of breast cancer patients and (D) the total number of bisphosphonate users in
each original study. Each circle represents an original study and the circle size is proportional to
the relative weight of the study using random-effects meta-regression analysis.
eFigure 9A
Mean Age of Participants in Each Original Study
Log
RR
54.5 63.8 70.0 73.4
0.00-0.04-0.08-0.12-0.16-0.20-0.24-0.28-0.32-0.36-0.40 60.0
Slope = 0.003P =0.65
0.080.14
eFigure 9B
Quality Score (%) of Each Original Study
Log
RR
73 75 78 81 86 89
0.00-0.04-0.08-0.12-0.16-0.20-0.24-0.28-0.32-0.36-0.40
Slope = 0.001P = 0.76
0.080.14
20 / 25
eFigure 9C
Total Number of Breast Cancer Patients in Each Original Study
Log
RR
5000 25000 37500 50000
0.00-0.04-0.08-0.12-0.16-0.20-0.24-0.28-0.32-0.36-0.40
12500200
Slope = 0.00001P = 0.11
0.080.14
eFigure 9D
Total Number of Bisphosphonate users in Each Original Study
Log
RR
200 12000 30000 42969 81329
0.00-0.04-0.08-0.12-0.16-0.20-0.24-0.28-0.32-0.36-0.40 60000
Slope = -0.00001P = 0.56
0.080.14
21 / 25
eFigure 10. Funnel plots of potential publication. White circles indicate individual studies; black
circles, trimmed studies after adjustment for potential publication bias; white diamonds, the
summarized effect estimates from the meta-analyses; black diamonds, the adjusted summary effect
estimates. The diagonal lines represent the 95% confidence interval of the summarized effect
estimate, which is indicated by the vertical solid line. RR, relative risk.
-1.0 -0.5 0.0 0.5 1.0
0.0
0.1
0.2
0.3
0.4
Stan
dard
Err
or
Log risk ratio
22 / 25
eFigure 11. Risk of bias of randomized controlled trials. (A) Judgements about each risk of bias
item for each included study; (B) Judgements about each risk of bias item presented as percentages
across all included studies.
A
B
23 / 25
APPENDICESAppendix A. PRISMA Checklist
Section/topic # Checklist item Reported on page # (or Brief description of how the criteria were handled in the meta-analysis)
TITLE
√ Title 1 Identify the report as a systematic review, meta-analysis, or both. Title page. Bisphosphonates and primary breast cancer risk: an updated systematic review and meta-analysis involving 963,995 women.
ABSTRACT
√ Structured summary 2Provide a structured summary including, as applicable: background; objectives; data sources; study eligibility criteria, participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implications of key findings; systematic review registration number.
Page 5-6. A structured abstract is provided on page 5-6.
INTRODUCTION
√ Rationale 3 Describe the rationale for the review in the context of what is already known. Page 7. Prevention of primary breast cancer (BCa) in women is of great public health importance. The existing results from observational epidemiologic studies focused on the association between bisphosphonates and primary BCa risk have been inconsistent.
√ Objectives 4 Provide an explicit statement of questions being addressed with reference to participants, interventions, comparisons, outcomes, and study design (PICOS).
Page 7. Does the use of bisphosphonates reduce the risk of primary breast cancer in women? The participants are women. The exposure or intervention is the use of any type of bisphosphonates. The outcome is the risk of developing primary breast cancer. The comparison is bisphosphonate users versus non-users. Study designs include cohort studies, case-control studies, and randomized controlled trials.
METHODS
√ Protocol and registration 5 Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if available, provide registration information including registration number. Page 8. This meta-analysis was registered on PROSPERO (CRD42014014901).
√ Eligibility criteria 6 Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale. Page 8-9. Detailed inclusion and exclusion criteria were described in the methods section.
√ Information sources 7 Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched.
Page 8. Databases: MEDLINE, EMBASE, ProQuest, and Web of Science;Time period: the first search - through August 15, 2016; the updated search – from August 15, 2016 to June 25, 2018. We hand-searched reference sections and citation lists of the retrieved literature for additional references. If an abstract was identified, we had tried to contact with the authors to access their data.
√ Search 8 Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated.
Page 8. We systematically searched several databases and conducted additional hand-searching to include all available studies. The detailed search terms are provided on page 8.
√ Study selection 9 State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis).
Page 9 and Figure 1. The eligibility determination for each study were performed independently by two authors (YP.L. and XS.Z.). Disagreements were resolved by discussion and consensus with a third author (HR.S.). The detailed process for study selection is provided as the flow chart (Figure 1).
√ Data collection process 10 Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators.
Page 9. The data extraction, and quality assessment for each study were performed independently by two authors (YP.L. and XS.Z.). Disagreements were resolved by discussion and consensus with a third author (HR.S.).
√ Data items 11 List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made.
Page 9. Standard electronic forms specifically created for the present study were used to record the variables, which were described in the main text. Also see eTable 1-3, 6 and 7 in the Supplementary Materials.
√ Risk of bias in individual studies 12 Describe methods used for assessing risk of bias of individual studies (including specification of whether this was
done at the study or outcome level), and how this information is to be used in any data synthesis. Page 9-10. A quality-scoring system was used for assessding risk of bias of observational studies; and the Jadad score and Cochrane risk of bias tool were used to assess the quality of RCTs.
√ Summary measures 13 State the principal summary measures (e.g., risk ratio, difference in means). Page 9. The principal summary measures include relative risks, odds ratios and hazard ratio with 95% confidence intervals.
√ Synthesis of results 14 Describe the methods of handling data and combining results of studies, if done, including measures of consistency (e.g., I2) for each meta-analysis.
Page 10-11. We preferentially used the maximally adjusted effect estimates in this meta-analysis. The 95% confidence intervals of summarized estimates were all provided. I2 values and P values for heterogeneity were reported. The summarized results with corresponding 95% confidence intervals of sensitivity analyses were provided. The confounding RR values and the corresponding P values were provided. The E values for the pooled estimates and the lower or upper limits were provided.
24 / 25
Section/topic # Checklist item Reported on page # (or Brief description of how the criteria were handled in the meta-analysis)
√ Risk of bias across studies 15 Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies).
Page 11. We used funnel plots for asymmetry, the Begg’s rank correlation and Egger’s linear regression tests, and the Duval and Tweedie’s trim and fill method to assess the potential publication bias.
√ Additional analyses 16 Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating which were pre-specified.
Page 10-11. We performed extensive sensitivity analyses, such as confounding RR, the E-value analysis, subgroup analyses, etc.
RESULTS
√ Study selection 17Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram.
Figure 1. Details of the literature search process are outlined in the flow chart (Figure 1).
√ Study characteristics 18For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations.
Page 12 in the main text and eTable 1-3, 6 and 7 in the Supplementary Materials.
√ Risk of bias within studies 19 Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12). Page 12 in the main text and eFigure 11 for RCTs and Appendix B for cohort and case-control studies in the Supplementary Materials.
√ Results of individual studies 20For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot.
Figure 2 in the main text and eFigure 1-8 in the Supplementary Materials.
√ Synthesis of results 21 Present results of each meta-analysis done, including confidence intervals and measures of consistency. Page 12-13. Figure 2 in the main text and eFigure 1-8 in the Supplementary Materials..
√ Risk of bias across studies 22 Present results of any assessment of risk of bias across studies (see Item 15). Page 14-15. eTable 8 and eFigure 10 in the Supplementary Materials.
√ Additional analysis 23 Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses, meta-regression [see Item 16]). Page 13-14. Table 1 in the main text; eTable 4-5 and eFigure 1, 2, 5-8 in the Supplementary Materials.
DISCUSSION
√ Summary of evidence 24Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy makers).
Page 15. We found a 12% risk reduction of BCa associated with bisphosphonates, which has strengthened the evidence for anticancer effects of bisphosphonates in the BCa primary prevention. However, according to the evidence synthesis, there is a general "indecisive" evidence for this association.
√ Limitations 25Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias).
Page 19-20. Detailed limitations of this study are provided in the Discussion section of the main text.
√ Conclusions 26 Provide a general interpretation of the results in the context of other evidence, and implications for future research.
Page 20. Our findings added to the body of evidence for the association between bisphosphonates and a decreased risk of primary breast cancer. However, we can not draw a decisive conclusion based on the current evidence. There is still a long way to go before the use of bisphosphonates as breast cancer chemoprevention strategy in routine practices.
FUNDING
√ Funding 27Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the systematic review.
Page 2. Funding sources are provided on the title page.
From: Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(7): e1000097. doi:10.1371/journal.pmed1000097.
25 / 25
Appendix B. Methodological quality score of the cohort and case-control studies
Study name
Score of methodological quality
1(0/4/8)
2(0/4/7)
3(0/4/7)
4(0/10)
5(0/8)
6(0/4)
7(0/4/6)
8(0/2)
9(0/6)
10(0/4/6)
11(0/4)
12(0/4)
13(0/8)
14(0/6)
15(0/4/9)
Selection bias 40
(1-5 items)
Misclassification bias 40
(6-13 items)
Confounding bias 15
(14-15 items)Total Reporting
score (%)*
Cohort studies
Chiang 2012Taiwan, China 0 7 7 10 8 4 6 0 6 6 4 0 8 6 4 32 34 10 76 80
Cardwell 2011United Kingdom 0 7 7 10 8 4 6 2 6 6 4 4 8 6 4 32 40 10 82 86
Vestergaard 2011Denmark 0 7 7 10 8 4 6 2 6 6 0 0 8 6 4 32 32 10 74 78
Chlebowski 2010USA 0 7 7 10 8 4 4 2 6 6 0 0 8 6 9 32 30 15 77 81
Lee 2012 Taiwan, China 0 7 7 10 8 0 6 0 6 6 4 0 8 0 4 32 30 4 66 69
Fournier 2017French 4 7 7 10 8 4 4 2 6 6 4 0 8 6 9 36 34 15 85 89
Kwan 2016USA 0 7 7 10 8 4 6 2 6 6 4 0 8 6 9 32 36 15 83 87
Korde 2018USA 4 7 7 10 8 0 6 2 6 6 4 0 8 6 4 36 32 10 78 82
Case-control studies
Rennert 2010Israel 0 0 7 10 8 4 6 2 6 4 4 0 8 6 9 25 34 15 74 78
Newcomb 2010USA 0 4 7 10 8 4 4 2 6 6 0 0 8 6 4 29 30 10 69 73
Vinogradova 2013 United Kingdom 0 0 7 10 8 4 6 2 6 6 4 0 8 6 4 25 36 10 71 75
Monsees 2011USA 4 4 7 10 8 0 6 2 6 6 4 0 8 6 9 33 32 15 80 84
*The quality score of each study was presented as a percentage of the maximum score, and studies with a score more than 50% were categorized as high-quality studies.
26 / 25