Effectiveness of preventive interventions in BRCA1/2 gene mutation carriers: A systematic review†
Version of Record online: 30 APR 2007
Copyright © 2007 Wiley-Liss, Inc.
International Journal of Cancer
Volume 121, Issue 2, pages 225–231, 15 July 2007
How to Cite
Bermejo-Pérez, M.J., Márquez-Calderón, S. and Llanos-Méndez, A. (2007), Effectiveness of preventive interventions in BRCA1/2 gene mutation carriers: A systematic review. Int. J. Cancer, 121: 225–231. doi: 10.1002/ijc.22817
This study has been carried out in accordance with the guidelines set by Andalusian Agency for Health Technology Assessment within the framework foreseen in the Quality Plan for the National Health System, within the collaboration agreement endorsed by Instituto Carlos III, an autonomous body of the Ministry of Health, and by the Fundación Progreso y Salud in Andalusia.
- Issue online: 18 MAY 2007
- Version of Record online: 30 APR 2007
- Manuscript Accepted: 3 APR 2007
- Manuscript Received: 23 FEB 2007
- BRCA 1 or BRCA 2;
- prophylactic mastectomy;
- prophylactic oophorectomy;
A systematic review of the literature was conducted to assess the outcomes of preventive interventions (prophylactic surgery, intensive cancer screening, and chemoprevention) in women who carry mutations in BRCA1/2 genes, in terms of reducing breast and gynaecological cancer incidence and/or mortality. A search for relevant articles published between 1996 and 2005 (inclusive) was run on Medline, Embase and other databases. From the 749 journal articles retrieved from this search strategy, 18 studies were eligible for this review (2 systematic reviews, 10 cohort studies and 6 case-control studies). The critical appraisal of the studies was performed by two independent reviewers with a list of ad hoc selected criteria. The synthesis of results was qualitative. Mastectomy and prophylactic gynaecological surgery (oophorectomy or salpingo-oophorectomy) reduced breast and gynaecological cancer incidence in carriers of BRCA mutations, by comparison to surveillance. However, all the studies presented flaws in internal and external validity, none of these preventive interventions is risk-free, and protection against breast and gynaecological cancer, as well as other cancers linked to BRCA mutations, is incomplete. No studies comparing surveillance programmes of varying intensity were found. Exposure to drugs (tamoxifen, and oral contraceptives) in women carrying BRCA mutations was assessed through a limited number of papers. All of these were case-control studies with prevalent cases and presented major methodological flaws. © 2007 Wiley-Liss, Inc.
Breast cancer is the most commonly diagnosed cancer after nonmelanoma skin cancer in women, and it is the second leading cause of cancer-related deaths. Although less common, ovarian cancer is associated with high morbidity and mortality rates and ranks as the fourth cause of all cancer-related deaths among Western women.1, 2
Most breast cancers (70–80%) occur in women with no family history of cancer (sporadic cancers). However, the remaining cases are all linked to a family history of cancer. The aggregation of breast cancer cases in a given family has also been related to an increased risk for ovarian cancer. Although different genes have been linked to hereditary cancer, most familial cases of breast and gynecological cancers (ovarian cancer, Fallopian tube cancer and carcinoma of the peritoneum) are associated with mutations in the BRCA1 and BRCA2 genes.
Germ line mutations in BRCA1/2 genes confer a higher risk of breast and ovarian cancer, clearly above the risk for the general population. However, the average magnitude of these risks is uncertain and may depend on context (intensity of family history, ethnic group, environmental factors, etc.). In women carrying BRCA1 mutations, the average cumulative risk of cancer by age 70 years ranged between 51 and 95% for breast cancer and between 22 and 66% for ovarian cancer. In BRCA 2-mutation carriers, the risks ranged between 33 and 95% for breast cancer and between 4 and 47% for ovarian cancer. The highest penetrance was detected in families with multiple cases of cancer.3
Some preventive interventions have been proposed for these women, such as prophylactic surgery (mastectomy and gynecological surgery), chemoprevention and intensive surveillance. However, the comparison of such interventions in terms of benefits and risks is subject to considerable uncertainty. For instance, at the beginning of 2005, there was very little information available on the effectiveness of these strategies in the population of interest (women carrying BRCA1/2 mutations). Despite this uncertainty, the high penetrance of such mutations and the increase in the number of laboratories performing gene testing point towards an increase in demand for such testing.4 There is a clear need for information, given the drastic and irreversible nature of preventive surgery and the likely adverse effects associated with the different preventive strategies.
It is in this context that the literature review was addressed, in order to answer the following research question:
What is the effectiveness of each preventive intervention strategy implemented in women carrying BRCA 1 or 2 mutations, in terms of reducing incidence and mortality from breast or gynecological cancer?
Once the articles had been selected in response to this research question, our approach was to describe other health outcomes (i.e. incidence on other kinds of cancer, all-cause mortality rates, physical and mental morbidity, quality of life, etc.) associated with any of the preventive intervention strategies.
Search strategy to pinpoint studies for this systematic review
A search for relevant articles was run from 1996 to 2005 (both included) on MEDLINE, EMBASE, Cochrane Library (The Cochrane Database of Systematic Reviews and Controlled Trials Register), ClinicalTrials.gov (http://clinicaltrial.gov/), the National Research Register of National Health Service (http://www.update-software.com/national/), the Centre for Reviews and Dissemination databases (International Network of Agencies for Health Technology Assessment, Database of Abstracts of Reviews of Effects, and NHS Economic Evaluation Database), and Web sites related to the study topic.
The MeSH used for the search on MEDLINE were “Genes-BRCA1,” “Genes-BRCA2,” “Breast-Neoplasms”/prevention-and-control, “Ovarian-Neoplasms”/prevention-and-control, “Fallopian-Tube-Neoplasms”/prevention-and-control, “Chemoprevention” and “Mass Screening.” The MeSH used for the search on EMBASE were “oncogene,” “breast-cancer”/prevention, “ovary-cancer”/prevention, “chemoprophylaxis” and “Mass Screening.” Several key free-text words were also used.
Last, a review was conducted based on the references provided in the selected articles.
Inclusion and exclusion criteria for the studies
The set of inclusion criteria for articles referred to study design (systematic reviews, clinical trials, observational studies including a control group), participants' characteristics (women carrying BRCA1/2 mutations), interventions (any strategy that was intended to ensure early detection of breast or gynecological cancers) and outcomes (incidence of breast or gynecological cancer or death from any cause).
When carriers of BRCA mutations only comprised a subgroup of the study population, the article was excluded if no separate data on cancer incidence and/or mortality in this subgroup could be extracted. Also excluded were any articles that were either not original or based outcomes on a hypothetical population, included a sample size of under 10 patients or provided insufficient information to ensure compliance with inclusion criteria.
Critical assessment of selected studies and data extraction
Two reviewers independently assessed the quality of studies. Any discrepancies between reviewers were resolved by a third reviewer.
A list of criteria was drawn up ad hoc for critical appraisal. It was based on the recommendations of the Critical Appraisal Skills Programme5 for cohort studies and case-control studies, and of the Reader's Guide to Critical Appraisal of Cohort Studies.6, 7, 8 To ensure specific assessment of any biases, we selected from the scientific literature a series of variables that may act as confounding factors or effect modifiers, either because they are related to the final endpoint (breast or gynecological cancer) and/or exposure (kind of preventive intervention). For each of these variables, the following information was extracted from the original studies; if it was used for defining women's inclusion or exclusion criteria, its distribution in each comparison group, and if it was used for adjusting the statistical analysis.
Information was taken from each of the selected articles by the same person. Basic information on the components of the research question was as follows: number of women carrying BRCA mutations in each group compared, together with their clinical characteristics, incidence of breast or gynecological cancer, mortality rate for each group and lenght of follow-up. The kind of information evidently also depended on study design (cohort or case-controlstudies).
To appraise internal and external validity, information was compiled from data provided in the articles on sociodemographic variables (mean age of the women, life styles, nationality, race, ethnic group) and clinical records (family and personal history of cancer, kind of BRCA mutation, gynecological history—age at menarche, number of children, use of exogenous hormones and any prior history of prophylactic surgery).
Synthesis of results
The appropriateness of conducting a meta-analysis was tested by means of the Q statistics, and Galbraith's and L'Abbé's graphs. The meta-analysis was inappropriate in the light of the existing statistical heterogeneity and differences among the populations in the original papers. Finally, a qualitative synthesis of results was performed, taking into account the critical appraisal of each original article.
About 1259 articles were identified using the above-mentioned search strategy. After removing duplicates, the number dropped to 749 articles. A further 677 of these was rejected on the basis of the information contained in the abstracts (297 were not original studies, 349 did not meet one or several of the inclusion criteria and sample size was under 10 patients in 31 articles). Of the remaining 72 articles, 54 were also rejected after reading the full-text version (31 did not meet the inclusion criteria set for this review; 18 failed to provide information on incidence or cancer-related mortality and a further 5 had a very limited number of women with BRCA mutations in the control group). Only 18 articles met all the inclusion criteria and were included in this review.
Description of the selected articles
Of the 18 articles, 2 were systematic reviews9, 10 while 16 were primary research studies. The 2 systematic reviews included studies that met the selection criteria for this review and others that did not (mainly because no results were provided separately on preventive interventions in women with BRCA gene mutations). Among the articles included in the systematic reviews, there were 5 that met our own inclusion criteria; all of these were among the 16 original studies selected previously.
The 16 primary research articles selected entail 22 statistical analyses (comparisons). These articles may be divided into 2 groups according to the study design:
- 1Cohort studies (n = 10).11–20 All 10 studies assessed preventive surgery interventions: 3 assessed prophylactic mastectomy against surveillance, 6 compared gynecological surgery and monitoring, and 1 compared 2 different gynecological surgery techniques.
- 2Case-control studies (n = 6).21–26 One of these studies assessed the effects of tubal ligation, one the effects of tamoxifen, and 4 addressed the effects of exposure to oral contraceptives (OC).
Some of the articles included (3 cohort studies12, 14, 15 and 4 case-control studies21–23, 25) stem from the same study. However, it was decided that all articles were to be included in the light of the distinct outcomes reported (for different comparisons, subpopulations of women or kind of outcome).
Studies comparing surveillance programmes of varying intensity could not be retrieved.
Methodological quality could be appraised in only 10 of the 16 articles selected (6 cohort studies and 4 case-controls) namely those articles with a study population comprising only women carrying BRCA gene mutations. As for the remaining 6 articles, only the outcomes for a part of the sample were selected (subgroup of women with BRCA mutation) as there was insufficient data available to appraise the similarity between study groups.17–25
Cohort studies on preventive surgery
The study population entailed women carrying BRCA 1 or BRCA 2 gene mutations who belonged to high-risk families.
The studies selected were observational and included prospective11, 12, 16 or retrospective follow-up.13–15 Methodological flaws were pinpointed in all the studies appraised, mainly due to selection bias. First, all the articles included self-selected comparison groups.11–16 Also, the process used to select women led to comparison groups with a different baseline risk for developing the endpoint. The following details should also be stressed: the inclusion of women in the control group after they have been diagnosed with cancer,11 the use of different criteria for defining the start point of follow-up in each of the comparison groups (date of preventive surgery in the intervention group and date when surveillance commenced in the control group).12–14 Also, some of the studies excluded women who had been diagnosed with carcinoma from the surgical specimen taken during prophylactic surgery.13, 16
Problems were also seen in the comparability between groups in most of the studies assessed. Among the variables with significantly different distribution between the intervention and control groups were year of birth,12, 13 age at menarche,12 use of exogenous hormones12, 14, 15 and undergoing preventive surgery concomitantly with the intervention assessed.11–13 Information was not always provided on patients lost to follow-up. In 2 of the articles providing these data, major differences in the percentage of women censored were noticed among the comparison groups.14, 15 No reasons were given for why patients were lost to follow-up.
Outcomes of preventive mastectomy.
Breast cancer incidence.
Prophylactic bilateral mastectomy (PBM) versus surveillance reduced the incidence of breast cancer in women carrying BRCA gene mutations with no previous history of cancer. Outcomes were taken from 2 articles including a total of 619 women. Hazard ratios ranged between 0 and 0.09 (statistically significant estimates), with 3- to 7-year follow-up11, 12 (Table I).
|Reference||Interventions appraised||Population (n)||Follow up (years)||Incident cancers1||Hazard ratio|
|Meijers11||Prophylactic bilateral mastectomy versus surveillance||76||63||2.9||3||0||8||0|
|Sprundel_A113||Prophylactic contralateral mastectomy versus surveillance||79||69||7.4||10.5||1||32||0.03 (0–0.19)2|
|Rebbeck14||Prophylactic bilateral oophorectomy versus surveillance||43||79||9.6||8.1||10||30||0.53 (0.33–0,84)3|
|Kauff_A216||Prophylactic bilateral salpingo-oophorectomy versus surveillance||69||62||0.8||1||3||8||0.32 (0.08–1.20)|
Prophylactic contralateral mastectomy (PCM) versus surveillance reduced the incidence of contralateral breast cancer in women with unilateral breast cancer carrying BRCA gene mutations. The results were taken from a single study including 148 women. The hazard ratio was 0.09 (statistically significant estimate) in just over 3 years' follow-up13 (Table I).
Death from any cause.
Only one article on PBM provides data on deaths occurring during the follow-up period (3 years), namely 4 among the 63 women in the control group (6%) and none in the 76 women in the intervention arm. Only one of the deaths was due to breast cancer.11
The only study on PCM reported 3 breast cancer deaths among the 79 women in the intervention arm (3.7% in 7.4 years of follow-up) and in 8 of the 69 controls (11.5% in 10.5 years follow-up). There were no statistically significant differences in survival of breast cancer seen in the comparison groups. Also, 3 more deaths were reported in the control group (2 due to ovarian cancer and 1 lung cancer case). There was only a significant reduction in breast cancer-related and all-cause deaths in the subgroup of women with PCM and bilateral oophorectomy (mostly intended as preventive therapy), compared with the women who had not undergone any intervention.13
Outcomes of gynecological surgery (oophorectomy and salpingo-oophorectomy).
Breast cancer incidence.
In the 3 articles examining this endpoint (with a total of 408 women), prophylactic gynecological surgery (PGS) versus surveillance brought about a reduction in breast cancer incidence. Outcomes differed in line with patients' personal history of breast cancer and follow-up time set for each study. For instance, reductions of around 50% in breast cancer risk were seen (hazard ratio between 0.47 and 0.53: primary breast cancer over the 11-year follow-up in women with no previous history14, 15); of 90% (hazard ratio 0.10: breast cancer recurrence within 3 years in women with a previous history17) and of 68% (hazard ratio 0.32: primary cancers and recurrences during 1-year follow-up; these results did not reach statistical significance16) (Table I).
Gynecological cancer incidence.
The effects of preventive gynecological surgery versus surveillance, in terms of reduction in gynecological cancer, were addressed in 4 studies, with different populations and follow-up periods (Table II). Very different outcomes were reported in these articles. For instance, in one, with a small sample size, no cancers were detected after a 5–7-year follow-up in any of the groups18 while in another, only one peritoneal cancer was reported in 152 women with a 2 and a half year follow-up.19 The 2 studies with the largest sample size (including 721 women) reported a 96% decrease in the incidence of gynecological cancers (hazard ratio 0.04) in 8 years,15 and an 85% decrease (hazard ratio 0.15, which was not statistically significant) in 2 years16; respectively.
|Reference||Interventions appraised||Population (n)||Follow-up (year)||Incident cancers1||Hazard ratio (95% CI)|
|Rebbeck_A115||PBO versus surveillance||259||292||8.2||8.8||Peritoneum 2||Ovary 58||0.04 (0.01–0.16)|
|Kauff_A116||PBSO versus surveillance||98||72||1.95||2.1||Peritoneum 1||Ovary 4||0.15 (0.02–1.31)|
|Olivier20||PBO versus PBSO||29||65||3.42||1||Peritoneum 3||0||–|
Apart from these 4 studies comparing preventive surgery with surveillance, a further study was retrieved that compared the effectiveness of 2 different techniques, namely prophylactic bilateral oophorectomy (PBO) and prophylactic bilateral salpingo oophorectomy (PBSO). Three peritoneal cancers were reported in women undergoing PBO after more than 3 years' follow-up and none among the women undergoing PBSO after 1-year follow-up20 (Table II). Also, in the group undergoing PBSO, more cancers were detected in the surgical specimens than in the PBO group (5 vs. 0—data not shown in the table on incident cancers).
All cause mortality.
Only one study on PGS provides data on survival.17 In our study, that included 36 women with BRCA 1 mutation and a history of breast cancer, there was a 67% 5-year survival rate in patients undergoing PBO versus 44% in the surveillance arm. Even though the difference in survival between the groups compared was significant (P = 0.01), no adjustment was made for possible confounding factors.
Safety of gynaecological surgery.
Only disaggregated data is provided for women with BRCA mutations in 1 study.16 Postsurgical complications occurred in 1%.
Qualitative synthesis of outcomes related to prophylactic surgery.
Compared to surveillance, prophylactic mastectomy and PGS lead to a reduction in the incidence of breast and gynecological cancer in women carrying BRCA gene mutations. The methodological flaws pinpointed, mainly selection bias, may lead to an overestimation of the effectiveness of surgical techniques (mainly in the studies on prohylactic mastectomy). However, it seems unlikely that such an overestimation could invalidate the outcome (protective effect of surgery).
It may not be concluded from the information provided by the selected studies that prophylactic surgery reduces mortality among women undergoing surgery compared to women who decide to continue with regular checkups. The works on prohylactic mastectomy showed no statistically significant differences as regards breast cancer-related mortality in the different comparison groups. The only study on PGS reporting a longer survival among women undergoing oophorectomy has both a small sample size and flaws in internal validity.
Case-control studies on drugs and Fallopian-tube ligation
The case-control studies retrieved provide information on the effects of exposure to tubal ligation and to drugs versus nonexposure. The study population was made up of women carrying BRCA 1 or 2 gene mutations and belonging to high-risk families. Women with the endpoint of interest (breast or ovarian cancer) were considered as cases, while the controls were comparable to the cases but without the endpoint of interest.
Methodological quality of the studies.
All the studies were case-control studies with prevalent cases.21–26 By including only surviving women curbs investigation into the relationship between exposure and the outcome of interest among the deceased women and those lost to follow-up when the study was performed. This circumstance leads to major selection bias.
A potential bias was also seen in poor differential classification according to exposure, mainly in the studies on drugs.22, 23, 26 As a result, in the cases of severe disease, the data on exposure is more readily remembered than in controls (in these articles, the method used to compile information entailed a survey conducted several years after the disease was diagnosed). This, together with the fact that no study compared the information to reliable sources allows us to rule out an overestimation of the potential risk of breast cancer associated with OC,23 or an underestimation of the potentially protective effect of OC against ovarian cancer26 and of tamoxifen against contralateral breast cancer.22
Effects of Fallopian tube ligation.
The evidence of a possible protective effect on ovarian cancer of tubal ligation is limited to a single case-control study with prevalent cases. The protective association was limited to women carrying BRCA 1 gene mutation (adjusted OR of 0.39; 95% CI: 0.22–0.70)21 (Table III).
|Reference||Exposure||Population (n)||Endpoint||Odds ratio|
|Cases||Controls||BRCA 1||BRCA 2||BRCA 1 or 2|
|Narod21||Fallopian-tube ligation||232||232||Ovarian cancer||0.39 (0.22–0.70)||1.19 (0.38–3.68)||–|
|Narod22||Tamoxifen||209||384||Contralateral breast cancer||0.38 (0.19–0.74)1||0.63 (0.20–1.50)1||0.50 (0.28–0.89)|
|Narod23||Oral contraceptives||1311||1311||Breast cancer||1.20 (1.02–1.40)||0.94 (0.72–1.24)||–|
|Whittemore26||147||304||Ovarian cancer||–||–||0.85 (0.53–1.4)|
Effects of tamoxifen.
The effects of tamoxifen in terms of prevention for contralateral breast cancer have only been appraised in one case-control study with prevalent cases, where a protective association of the drug was reported. However, this only reached statistical significance in the subgroup of women bearing mutations on BRCA 1 genes (nonadjusted OR of 0.38; 95% CI: 0.19–0.74)22 (Table III).
Effects of oral contraceptives.
The relationship between OC and breast cancer risk was appraised in 2 case-control studies with prevalent cases.23, 24 A positive association was only found in one of these, although of minor magnitude (OR of 1.20; 95% CI: 1.02–1.40)23 (Table III). The relationship between OC and ovarian cancer risk was also appraised in 2 case-control studies with prevalent cases.25, 26 A negative association was only seen in one of these (OR of 0.40; 95% CI: 0.2–0.7), precisely in the study with a lower number of controls25 (Table III).
Qualitative synthesis of effects related to exposure to Fallopian-tube ligation and drugs.
Given the small number of studies appraising each exposure factor and the flaws in validity detected (bias selection and poor classification according to exposure) the evidence on the effect of tubal ligation, tamoxifen and OC on breast or ovarian cancer incidence in women bearing BRCA gene mutations is very limited.
The studies included in this review show that there is insufficient evidence on both chemoprevention and benefits of intensive surveillance in women carrying BRCA mutations. An update of the only study on tamoxifen and contralateral breast cancer appraised in this review22 has been recently published. Our study show a protective association not only in women carrying BRCA 1 gene mutations, as reported in the previous paper, but also in women carriers of BRCA2 gene mutations (adjusted OR of 0.42; 0.17–1.02).27 However, the study design (case-control with prevalent cases) still poses a limitation on the reliability of the outcomes reported.
After examining the studies included in this review, we can state that mastectomy and PGS reduced breast and gynecological cancer incidence in carriers of BRCA mutations. A further study with sound methodological quality (which has not been included in this review given its recent publication date) also provides information on this topic. Here, a reduction in the risk for gynecological cancer was seen in the group undergoing PBSO compared with the control group (HR:0.20; 95% CI:0.07–0.58).28
On the other hand, no increase in survival was proven in the selected studies (up to 2005). However, another one published later shows a reduction in overall mortality (HR: 0.24; 95% CI: 0.08–0.71), breast-cancer-specific mortality (HR: 0.10; 95% CI: 0.02–0.71), and ovarian-cancer-specific mortality (HR: 0.05; 95% CI: 0.01–0.46) for women who had BPSO compared to those who did not.29
However, all these results must be interpreted with caution. First, more evidence supporting reduction in mortality is required, taking into account that there is only one study showing the outcome. Second, all the studies appraised present biases and flaws in internal and external validity. Finally, risk cannot be ruled out for any of these preventive interventions. Likewise, protection against breast and gynecological cancer, as well as other cancers linked to BRCA mutations is incomplete.
In a systematic review of the literature, we must always bear in mind that results are based on the findings of completed studies. As a result, the review is subjected to limitations in quality and number of articles. Also, the articles retrieved from a bibliography search may entail a biased sample of all the studies conducted, given that those articles with outcomes supporting the initial hypothesis tend to be published more frequently.30
Interpretation of results
It must also be borne in mind that the studies present their results in relative terms (hazard ratio). Although from the technical standpoint, this is a correct approach and it also tends to magnify the final outcome. When interventions with a potentially protective effect are under appraisal, a more realistic data set may be provided by estimating absolute risk reduction for exposed patients. By way of example, in one study, the relative reduction in breast cancer risk linked to PBM was 100% but absolute risk reduction was only 12%.11
Another major issue when appraising the effectiveness of interventions of this kind involves understanding the staging of cancer diagnosed in the group that is not undergoing a preventive intervention. Most works make no distinction between invasive carcinoma and carcinoma in situ. This is of vital importance given that both conditions are very distinct entities in terms of prognosis and it would provide relevant information on the real risk of disease-related death when the decision taken is to continue with regular checkups.
It should also be pointed out that neither prophylactic mastectomy nor PGS completely eradicate the risk of breast or gynecological cancer.12, 28, 31 Besides, the risk of other tumors also linked to BRCA mutations, such as stomach, pancreatic or colon cancers, amongst others, still remains.32
The risks of surgery itself and anesthesia must also always be taken into account when addressing any surgical technique, even in healthy subjects. In the studies appraised, no information is provided on complications in prophylactic mastectomy. However, the literature reports postsurgical complication rates of up to 30.4% in women undergoing mastectomy and immediate reconstruction.33 Only one of the studies selected for this appraisal reported complications associated with PGS of under 1%.16 These results are similar to those seen in other studies on complications after laparoscopic gynecological surgery,34 although this is lower than the complication rate associated with open surgery (17%).35 Excision of the ovaries also leads to sterility and premature menopause with its associated symptoms, which are usually relieved with drugs that cannot claim to be harmless.36
As for psychological morbidity, one review provided information on psychological issues seen in high-risk women undergoing prophylactic mastectomy.9 According to our study, most women stated that they were satisfied with their decision to undergo mastectomy, although this satisfaction was inversely related to surgical complications and a poor cosmetic outcome. Body image and feelings of femininity were the most salient psychosocial issues appraised. Contradictory results as regards the psychological impact of PGS in high-risk women are reported in the literature. For instance, in one study, PBO reduced concern over ovarian cancer and most participants stated a high degree of satisfaction with the decision taken.37 In another work, however, women who had undergone the procedure had worse social function scales and emotional role scores (Short Form-36 Health Status Questionnaire), they had fewer menopausal symptoms, while no significant differences were reported between these women and those who decided to continue with regular checkups in terms of anxiety over cancer and sexual function.38
External validity of the results
The effectiveness of preventive strategies is directly related to the penetrance of BRCA gene mutations in the women investigated. The external validity of the results is limited when the penetrance of certain BRCA gene mutations in different populations is unknown.
Besides the scant information on the population base in the articles selected, the high incidence of cancers diagnosed in the control group is striking in most articles. These data suggest that the control group cohorts were made up of women at very high risk for cancer that would be comparable to information published in the literature for lifetime penetrance of BRCA 1 or 2 gene mutations in families where there are multiple cases of cancer.3 Although it is true to say that most women undergoing genetic testing have more than one affected relative, there are also people who also undergo testing based on a weak family history or early diagnosis of the disease.39 This issue should be taken into account, given that the effectiveness of the interventions appraised may be lower in women with a lower risk for cancer.
Contributions of this review and implications of the results
Two systematic reviews were published during the drafting of this review. Amongst other topics, these reviews examined the effectiveness of preventive interventions to reduce the incidence and mortality of breast and ovarian cancer in women with BRCA mutations. This points to the need for a study on these characteristics at the design stage of our review. In the work by the Canadian Coordinating Office for Health Technology Assessment, the research covered up to July 2004,40 while the other study commissioned by the U. S. Preventive Services Task Force included publications up to October of the same year.41 Both reviews included articles with a population made up of women at high risk for breast and gynecological cancer, determined on the basis of detected BRCA mutations, family history or results from statistical models. Although the reviews reported conclusions as attributable to women carrying BRCA mutations only, in reality the results used were from some original articles where the populations included women with and without mutations. Similar conclusions have been drawn in spite of these differences among the 3 reviews and the fact that our study included a further year of publications.
Compared to the above-mentioned works, our review would contribute a greater number of selected studies on preventive interventions, which present disaggregated results for women with BRCA mutations; a deeper analysis of potential biases, and of their repercussions on the final results, as well as greater power in the external validity of the conclusions (selection of results exclusively derived from the population of interest).
Bearing in mind the implications involved in undergoing any of the preventive strategies, it seems clear that the population who will be offered the option of both genetic testing and preventive alternatives must be carefully selected. Once the appropriate population has been selected, the affected women themselves should take the final decision. Along these lines, genetic testing performed by trained professionals should meet all needs for information. If we also bear in mind the uncertainty of certain prophylactic strategies and the different values held by each individual, all the information should preferably be provided in a standardized fashion using decision-making tools.
Last, it also seems that the evidence on the effects of the preventive interventions requires enhancement through well-designed studies capable of providing more thorough information on the main variables that may assist in the appraisal of internal and external validity.
- 1BRCA1/2 and the prevention of breast cancer. In: KhouryMJ, LittleJ, BurkeW, editors. Human genome epidemiology. Oxford: Oxford University Press; 2004. 451–74..
- 2American cancer society: cancer facts and figures 2005. Atlanta, GA: American Cancer Society, 2004 [consulted on May 6, 2006]. Available from: http://seer.cancer.gov/csr/1975_2002/results_single/sect_01_table.01.pdf.
- 3Contribution of BRCA 1/2 mutation testing to risk assessment for susceptibility to breast and ovarian cancer. Montreal: AETMIS, 2006. Summary Report., , .
- 5National Health Service. Critical Appraisal Skills Programme [Internet]. UK: CASP, 2004. [consulted on July 6, 2005]. Available from: http://www.phru.nhs.uk/casp/appraisa.htm.
- 9Prophylactic mastectomy for the prevention of breast cancer. Cochrane Database Syst Rev 2004; 18: CD002748. DOI: 10.1002/14651858.CD002748.pub2., , , .
- 32BRCA1 and BRCA 2 pathway and the risk of cancers other than breast or ovarian. MedGenMed 2005; 7: 60..
- 40BRCA1 and BRCA2 predictive genetic testing for breast and ovarian cancers: a systematic review of clinical evidence, Tech. Rep. No. 66. Ottawa: Canadian Coordinating Office for Health Technology Assessment, 2006., , , , , , , .