BREAST

Outcomes Article

Surgical Determinants of Patient-Reported

Outcomes following Postmastectomy
Reconstruction in Women with Breast Cancer
Ranjeet Jeevan, M.R.C.S. Background: This national prospective cohort study compared the patient-report-
John P. Browne, Ph.D. ed outcomes of breast cancer patients undergoing postmastectomy autologous re-
Carmel Gulliver-Clarke, Ph.D. construction to those who had breast implants, in terms of aesthetic appearance;
Jerome Pereira, F.R.C.S. levels of psychological, physical, and sexual well-being; and overall satisfaction.
Christopher M. Caddy, Methods: Of 5063 women who underwent immediate reconstruction (n = 3349)
F.R.C.S.(Plast.) or delayed reconstruction (n = 1714) between January 1, 2008, and March 31,
Jan H. P. van der Meulen, 2009, in England, 2923 women who gave informed consent were sent validated,
Ph.D. procedure-specific, 18-month follow-up questionnaires. Outcome scale scores
David A. Cromwell, Ph.D. ranged from 0 (poor) to 100 (excellent); multiple linear regression was used
Worthing, London, Prescot, Merseyside, to adjust scores for patient and treatment characteristics.
West Sussex, Great Yarmouth, Norfolk, Results: Two thousand two hundred eighty-nine women (78 percent) returned
and Sheffield, United Kingdom; and completed questionnaires (immediate reconstruction, n = 1528; delayed recon-
Cork, Ireland struction, n = 761). For immediate reconstruction, mean overall satisfaction
scores for the various techniques ranged from 67 to 85 (median, 67 to 93). For
delayed reconstruction, mean overall satisfaction scores ranged from 70 to 85
(median, 75 to 100). For both groups, similar gradients were observed for the
other outcome scales across techniques. Reconstruction using patients’ own
tissues tended to have higher mean adjusted scores compared with those tech-
niques using implants alone (p < 0.0001 for aesthetic appearance, psychologi-
cal well-being, sexual well-being, and satisfaction with outcomes for immediate
and delayed reconstruction groups).
Conclusions: Women who underwent autologous reconstruction tended to re-
port greater satisfaction than those who underwent implant reconstruction.
These results can inform patients of the anticipated outcomes of their selected
surgery, but further research is required to confirm whether autologous re-
construction is superior in general. (Plast. Reconstr. Surg. 139: 1036e, 2017.)
CLINICAL QUESTION/LEVEL OF EVIDENCE: Therapeutic, II.

W
omen who undergo mastectomy surgery ways.1 All of the options aim to address the aes-
for breast cancer can have their breast thetic impact of mastectomy and may also improve
mound reconstructed in a variety of patients’ physical, emotional, and sexual well-
being.2 There has been much debate regarding
From the Clinical Effectiveness Unit, The Royal College of the best way to assess the effectiveness of differ-
Surgeons of England, the Health Services Research Unit, ent types of reconstructive procedures.3,4 Interna-
Department of Public Health & Policy, London School of tionally, the consensus view has shifted toward the
Hygiene & Tropical Medicine, St Helens and Knowsley
Teaching Hospitals NHS Trust, Western Sussex Hospitals
NHS Foundation Trust, Worthing Hospital, James Paget Disclosure: There was no funding source for this
University Hospitals NHS Foundation Trust, University of study. There was no support from any organization for
East Anglia, Norwich Research Park, and Sheffield Teach- the submitted work; there were no financial relation-
ing Hospitals NHS Foundation Trust, Royal Hallamshire ships with any organizations that might have an inter-
Hospital; and the Department of Epidemiology and Public est in the submitted work in the previous three years;
Health, University College Cork. there were no other relationships or activities that could
Received for publication May 18, 2016; accepted October appear to have influenced the submitted work. None of
5, 2016. the authors has a financial interest in any of the prod-
Copyright © 2017 by the American Society of Plastic Surgeons ucts or devices mentioned in this article.
DOI: 10.1097/PRS.0000000000003236

1036e www.PRSJournal.com
Copyright © 2017 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Volume 139, Number 5 • Breast Reconstruction Patient Outcomes

use of patient-reported outcome measures rather process. Approval to prospectively collect patient
than anthropomorphic measurements or mea- data for analysis and reporting was obtained from
surement scales scored by clinicians.2 This is in the Patient Information Advisory Group under
part because previous studies have found little cor- Section 60 of the Health and Social Care Act 2001.
relation between clinician-reported and patient-
reported aesthetic satisfaction following such Reconstructive Procedure Types
surgery and partly because of a recognition that We compared six different surgical tech-
the patient should be the ultimate judge of subjec- niques: tissue expander or breast implant alone;
tive outcomes such as aesthetic appearance.5,6 pedicled latissimus dorsi flap combined with
Numerous studies have evaluated patient- an expander or implant (latissimus dorsi with
reported outcomes following different types of implant/expander); latissimus dorsi flap alone
breast reconstruction in populations of between 45 (autologous latissimus dorsi); pedicled trans-
and 2328 women.6–15 However, these studies dem- fer of a transverse rectus abdominis myocutane-
onstrated methodologic weaknesses in one or more ous flap (TRAM); free tissue transfer of this flap
of the following areas: inadequate length of follow- (free TRAM); or a free tissue transfer of a similar
up; a failure to include immediate reconstruction flap without the underlying muscle, based on the
and delayed reconstruction patients, the full range deep inferior epigastric artery perforator (DIEP)
of reconstructive techniques, and all eligible cen- or superficial inferior epigastric artery (SIEA).
ters; lack of a prospective design; a lack of case-mix The remaining patients had a flap of skin and
adjustment; a lack of validated and surgery-specific fat with or without muscle taken from the upper
scales; and a failure to formally calculate outcome or lower buttock or inner thigh regions (super-
scores for these scales to enable valid comparisons. ficial gluteal artery perforator, inferior gluteal
We undertook a national prospective cohort artery perforator, or transverse upper gracilis) but
study that examined patient-reported outcomes were excluded from our analyses because of the
after mastectomy and breast reconstruction sur- extremely small numbers enrolled.
gery for women treated in England, Scotland, and
Wales.16 The study collected data from women Clinical Data Collection
undergoing immediate or delayed reconstruction A range of data items were recorded by clinicians
procedures using a set of validated surgery-spe- for each patient. These included details of surgical
cific scales that have been used widely to evaluate procedures, patient clinical and sociodemographic
the outcomes of breast reconstruction.2,17 In this characteristics, and consent status. The full data set
article, we compare the outcomes of six different is available at https://www.rcseng.ac.uk/surgeons/
reconstructive techniques. research/ceu/copy_of_docs.html.

PATIENTS AND METHODS Patient Questionnaires

The postreconstruction questionnaires addressed
Treatment Sites and Inclusion Criteria patient satisfaction with their breast area appear-
During the 15-month recruitment period from ance (16 items, with an additional two for implant
January 1, 2008, to March 31, 2009, all 150 English patients), which we believe to be synonymous with
National Health Service acute hospital trusts pro- aesthetic outcome from the patient’s perspective,
viding acute breast cancer treatment, six National along with emotional well-being (10 items), physical
Health Service acute trusts in Wales and Scotland, well-being (16 items), sexual well-being (six items),
and 114 independent hospitals in England partici- and the outcomes of surgery (seven items). The
pated in the study. Treatment sites were asked to scales were derived from the BREAST-Q,2 and were
prospectively collect and record clinical data on pretested with English breast cancer patients by
women aged 16 years and older with a diagnosis the authors before their use to ensure that there
of invasive carcinoma of the breast or ductal carci- were no issues with language, comprehension, or
noma in situ undergoing mastectomy surgery with acceptability. Copies of the full questionnaires are
immediate or primary delayed reconstruction fol- available at https://www.rcseng.ac.uk/surgeons/
lowing a previous mastectomy, and to obtain writ- research/ceu/copy_of_docs.html.
ten consent from eligible women to allow them
to be sent follow-up questionnaires. National Can- Questionnaire Administration and Collection
cer Audits are exempt from the United Kingdom Clinicians were asked to obtain consent from
National Research Ethics Committee approval any eligible women at the time of surgery to allow

1037e
Copyright © 2017 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Plastic and Reconstructive Surgery • May 2017

follow-up questionnaire administration. The ques- rank are produced for each geographic area
tionnaires were sent to patients who had given (super output area, a small local population of a
consent at their home address 18 months after few thousand people).
surgery by a coordinating team of researchers The Eastern Cooperative Oncology Group
that did not include the treating hospitals or clini- functional status categorical scoring system is an
cians, once the team had confirmed the patient American measure used internationally to mea-
was still alive by cross-checking their details against sure the functional status of patients undergoing
mortality data held by the National Strategic Trac- cancer treatments. It is also known as the World
ing Service. A prepaid envelope was enclosed to Health Organization or Zubrod score.
facilitate return of the completed questionnaire. The models were then used to produce
Questionnaires were marked only with a unique adjusted means and confidence intervals for each
numeric patient identifier. Nonrespondents were scale score, by procedure type, which demon-
sent a single reminder letter and an additional strated the effect size. Finally, the Wald test was
copy of the questionnaire at a 5-week interval. undertaken for the type of reconstructive proce-
dure to examine the heterogeneity of outcomes
Statistical Analysis and Multiple Linear across different types of reconstruction and deter-
Regression Model mine whether or not the differences in means
Patients’ responses were entered into a data- were attributable to chance alone.
base and then transferred to a bespoke Q-Score All statistical tests were two-sided, and values
software package to calculate scores for each of p < 0.05 were considered to represent a statisti-
BREAST-Q scale, with 0 being the lowest and 100 cally significant result. All statistical analyses were
being the highest possible scores.2 Patient scores undertaken using STATA/MP 14 (StataCorp, Col-
were linked to their associated clinical data using lege Station, Texas) and Microsoft Excel (Micro-
their unique numeric identifier. soft Corp., Redmond, Wash.) software.
Separate linear multiple regression models
were developed for the immediate and delayed RESULTS
reconstruction patient groups, and were used to
predict the outcome scores for each BREAST- Patient Population
Q scale (dependent variables) based on patient Within the study cohort, 3349 patients under-
characteristics and reconstructive procedure type. went immediate reconstruction. Of these, 43 did
The preliminary models included only patients not have their consent status for the follow-up
with complete outcome and case-mix data, and study recorded, 35 were deemed incapable of
were constructed using a backward stepwise pro- completing a written questionnaire in English,
cess with variables dropped from the models if the and 1148 were not asked to participate because
strength of their association with an outcome was of difficulties with the recruitment process at
weak (Wald test for variable inclusion, p < 0.05). some hospitals. Of the remaining 2123 women,
Variables were included in continuous or categor- 1939 gave their consent and 1528 returned a com-
ical format depending on their type. The case- pleted questionnaire. After excluding 144 women
mix variables included sociodemographic items for whom complete case-mix data were not avail-
(i.e., age, a geographically assigned measure of able, responses from 1384 were included in our
socioeconomic deprivation, and hospital of treat- final analyses. Questionnaires were completed
ment), patient factors known to affect subsequent 586 days after surgery, on average, with an inter-
morbidity (i.e., smoking status, body mass index, quartile range of 30 days.
diabetes status, American Society of Anesthesiolo- There were 1714 patients who underwent
gists grade, and Eastern Cooperative Oncology delayed reconstruction. Of these, nine did not
Group performance status), and tumor character- have their consent status for the follow-up study
istics (i.e., invasive status and Nottingham Prog- recorded, six were deemed incapable of complet-
nostic Index).18–21 ing a written questionnaire in English, and 609
Deprivation in England is measured by the were not asked to participate. Of the remaining
government using seven distinct domains or indi- 1090 asked to participate, 984 gave their consent
cators of poverty (i.e., income; employment; edu- and 761 retuned questionnaires. A further 28
cation, skills, and training; health; crime; barriers patients were excluded because of incomplete
to housing and services; and living environment) case-mix data; this left 733 for our final analy-
that are weighted before a deprivation score and ses. The mean length of time between a patient’s

1038e
Copyright © 2017 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Volume 139, Number 5 • Breast Reconstruction Patient Outcomes

mastectomy and their subsequent delayed recon- ethnicity, Eastern Cooperative Oncology Group
struction procedure was, on average, 2.8 years score, and diabetes status for those undergoing
(range, 0 to 32 years). There were small differ- delayed reconstruction. The other character-
ences in the mean time to delayed reconstruction istics examined (including age, radiotherapy,
by procedure type, varying from 2.4 years (latis- chemotherapy, smoking status, and body mass
simus dorsi flap with implant) to 2.9 (free flap) index) were not independently associated with
and 3.0 years (implant/expander-only). Question- outcomes and were not therefore included in
naires were completed 582 days after surgery, on the final models. The immediate reconstruc-
average, with an interquartile range of 26 days. tion patients were of similar age and functional
Table 1 summarizes the consent and response status to those undergoing delayed reconstruc-
rates for the different reconstruction techniques. tion, but were less likely to be from a white eth-
nic group or the most deprived quintile, or to
Patient Characteristics across the Surgical have a high American Society of Anesthesiolo-
Groups gists grade or diabetes. Immediate reconstruc-
Table 2 displays the sociodemographic tion patients were also much more likely to have
and clinical characteristics of the women who a low tumor burden (ductal carcinoma in situ
underwent each type of procedure. Patients or low-risk invasive).
undergoing different procedures were gener-
ally similar. Those characteristics that were sig- BREAST-Q Scores for Each Procedure
nificantly associated with the outcomes under Figure 1 presents the immediate and delayed
investigation were included in the two separate reconstruction procedure-specific unadjusted
multiple linear regression models used subse- mean scores for each of the five BREAST-Q scales,
quently to derive final case-mix adjusted scores along with the associated 95 percent confidence
for the respective patient groups. During the intervals. With respect to breast area appearance,
model development process, the characteris- in the immediate reconstruction group, women
tics significantly associated with outcomes were who underwent a pedicled TRAM flap proce-
ethnicity, deprivation level, American Society dure had the highest outcome scores. However,
of Anesthesiologists grade, and tumor burden the sample size for this group was relatively small
for immediate reconstruction patients; and (n = 34) and thus the precision associated with

Table 1. Patient Recruitment and Response Rates Associated with the Different Reconstructive Procedure
Groups, in Both Immediate and Delayed Reconstruction Settings
LD with
Expander/ Implant or Autologous Pedicled Free DIEP/
Implant (%) Expander (%) LD (%) TRAM (%) TRAM (%) SIEA (%) Overall
Immediate reconstruction
No. of patients in study 1242 729 850 82 92 354 3349
No. of patients asked to
participate 759 (61.1) 453 (62.1) 594 (69.9) 52 (63.4) 60 (65.2) 222 (62.7) 2140
No. of patients who gave
their consent 681 (89.7) 414 (91.4) 520 (87.5) 41 (78.8) 57 (95.0) 204 (91.9) 1917
No. of patients who
returned a questionnaire 522 (76.7) 330 (79.7) 430 (82.7) 37 (90.2) 44 (77.2) 165 (80.9) 1528
No. of patients who
returned a questionnaire
and had complete clinical
data 475 284 402 34 36 153 1384
Delayed reconstruction
No. of patients in study 280 434 356 90 191 363 1714
No. of patients asked to
participate 169 (60.4) 253 (58.3) 251 (70.5) 58 (64.4) 115 (60.2) 248 (68.3) 1094
No. of patients who gave
their consent 149 (88.2) 233 (92.1) 216 (86.1) 49 (84.5) 104 (90.4) 222 (89.5) 973
No. of patients who
returned a questionnaire 109 (73.2) 183 (78.5) 166 (76.9) 35 (71.4) 85 (81.7) 183 (82.4) 761
No. of patients who
returned a questionnaire
and had complete clinical
data 104 176 160 35 83 175 733
LD, latissimus dorsi.

1039e
Copyright © 2017 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Plastic and Reconstructive Surgery • May 2017

Table 2. Patient Characteristics by Reconstructive Procedure Type

LD Flap with
Expander/ Implant or Autologous Pedicled Free TRAM DIEP/
Implant Expander LD Flap TRAM Flap Flap SIEA Flap Total
Immediate reconstruction
No. of responders 475 284 402 34 36 153 1,384
Age, yr
Median 51 51 53 52 51.5 51 51
IQR 15 14 13 12 14.5 12 14
Ethnicity, %
White 85.1 84.5 83.3 97.1 88.9 81.1 84.4
Nonwhite 3.4 4.6 4.0 0.0 0.0 9.2 4.3
Not known/not stated 11.6 10.9 12.7 2.9 11.1 9.8 11.3
Index of multiple
deprivation quintile, %
1 (least deprived) 28.8 34.5 28.1 32.4 36.1 33.3 30.6
2 20.0 25.4 26.4 17.7 33.3 23.5 23.6
3 22.1 16.9 21.9 8.8 22.2 21.6 20.6
4 14.1 13.7 12.7 29.4 2.8 12.4 13.5
5 (most deprived) 15.0 9.5 11.0 11.8 5.6 9.2 11.7
ASA grade, %
1 69.1 76.4 73.1 70.6 72.2 69.9 72.0
2 28.0 23.2 26.1 29.4 25.0 30.1 26.7
3 or 4 3.0 0.4 0.8 0.0 2.8 0.0 1.4
ECOG status, %
0 90.9 93.3 92.0 91.2 94.4 96.1 92.4
1 7.0 6.4 7.5 8.8 5.6 3.3 6.6
≥2 2.1 0.4 0.5 0.0 0.0 0.7 1.0
Diabetic, % 1.7 1.1 1.5 2.9 2.8 1.3 1.5
Tumor burden, %
DCIS 26.5 27.8 32.3 41.2 36.1 26.8 29.1
Low NPI 36.2 37.7 30.1 23.5 27.8 26.8 33.2
Intermediate NPI 24.0 22.9 21.4 26.5 16.7 28.8 23.4
High NPI 13.3 11.6 16.2 8.8 19.4 17.7 14.3
Delayed reconstruction
No. of responders 104 176 160 35 83 175 733
Age, yr
Median 55 50.5 52 52 50 51 51
IQR 11.5 14 14 14 11 13 14
Ethnicity, %
White 92.3 90.9 89.4 88.6 79.5 91.4 89.5
Nonwhite 1.9 2.3 3.1 0.0 6.0 2.9 2.9
Not known/not stated 5.8 6.8 7.5 11.4 14.5 5.7 7.6
Index of multiple
deprivation quintile, %
1 (least deprived) 24.5 23.2 21.2 34.4 35.0 30.2 26.5
2 28.4 22.0 23.8 18.8 21.3 25.0 23.8
3 18.6 20.8 19.9 12.5 17.5 19.2 19.2
4 15.7 19.1 14.6 15.6 17.5 12.2 15.6
5 (most deprived) 12.8 14.9 20.5 18.8 8.8 13.4 14.9
ASA grade, %
1 66.4 59.1 70.0 77.1 51.8 68.6 64.8
2 28.9 37.5 28.8 22.9 44.6 30.9 32.9
3 or 4 4.8 3.4 1.3 0.0 3.6 0.6 2.3
ECOG status, %
0 88.5 90.3 91.3 91.4 92.8 95.4 91.8
1 7.7 9.7 7.5 8.6 7.2 4.6 7.4
≥2 3.9 0.0 1.3 0.0 0.0 0.0 0.8
Diabetic, % 1.9 2.3 3.8 2.9 1.2 3.4 2.7
Tumor burden, %
DCIS 15.8 8.5 19.2 6.9 8.2 11.7 12.6
Low NPI 37.6 25.5 19.9 31.0 30.1 29.2 31.4
Intermediate NPI 26.7 38.2 29.8 31.0 30.1 29.2 31.4
High NPI 19.8 27.9 31.1 31.0 37.0 32.5 29.6
LD, latissimus dorsi; IQR, interquartile range; ASA, American Society of Anesthesiologists; ECOG, Eastern Cooperative Oncology Group; DCIS,
ductal carcinoma in situ; NPI, Nottingham Prognostic Index.

this estimate is relatively poor. The other patient scores. The lowest breast area appearance scores
groups who underwent a flap reconstruction were associated with patients who underwent
(with or without implant) also had relatively high expander or implant-only reconstruction.

1040e
Copyright © 2017 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Volume 139, Number 5 • Breast Reconstruction Patient Outcomes

Fig. 1. Forest plot of unadjusted mean scores for each BREAST-Q scale following immediate and delayed reconstruction, with asso-
ciated 95 percent confidence intervals. LD, latissimus dorsi; Ped, pedicle; IR, immediate reconstruction; DR, delayed reconstruction.

In the delayed reconstruction group, free procedure, with mean overall satisfaction scores
TRAM flaps, DIEP flaps, and then the other types of 71.1 (95 percent CI, 69.8 to 72.4) for immedi-
of flap-based reconstruction (with or without ate reconstruction procedures and 79.3 (95 per-
implant) were associated with the highest breast cent CI, 77.7 to 81.0) for delayed reconstruction
area appearance scores. Again, the expander or procedures. Patients who underwent autologous
implant-only patient group had the lowest scores. reconstruction reported higher scores than those
After adjustment for patient characteristics, women who had an expander or implant alone, in
there were persistent differences in the outcomes both the immediate and delayed reconstruction
achieved by the procedure groups, with the settings, across four of the five scales. For immedi-
exception of physical well-being. Tables 3 and 4 ate procedures, the size of these differences was
show the adjusted differences in the mean scores typically between 1 and 7 points. For delayed pro-
for each scale, using the expander/implant group cedures, the difference between the scores tended
as the reference category. In both the immediate to be between 7 and 15 points.22 To interpret the
and delayed reconstruction settings, all flap-based magnitude of these differences, it is useful to
procedures resulted in higher scores on all scales compare them to a cohort of patients who under-
other than physical well-being. The only excep- went mastectomy alone over the same period
tion to this pattern was for the immediate recon- at the same group of hospitals.16 These patients
struction pedicled TRAM flap group. recorded a mean score of 56 on the breast area
appearance scale, which is similar to the mean
DISCUSSION unadjusted score recorded for implant-only
patients recorded in this study but between 8 and
Key Findings 21 points lower than that recorded for patients
Women who underwent reconstruction undergoing autologous immediate reconstruc-
tended to be satisfied with the results of the tion procedures. A similar pattern was seen for

1041e
Copyright © 2017 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Plastic and Reconstructive Surgery • May 2017

Table 3. Adjusted Mean Difference in BREAST-Q Scores by Type of Reconstruction, with Expander/Implant as
Reference, for Each Immediate Reconstruction Procedure Type
Breast Area Emotional Physical Sexual Satisfaction with
Appearance Well-being Well-being Well-being Outcome
Reconstruction Mean Mean Mean Mean Mean
Type Difference 95% CI Difference 95% CI Difference 95% CI Difference 95% CI Difference 95% CI
Expander/implant Reference
LD flap with implant
or expander 8.2 5.3–11.1 7.7 4.6–10.8 0.7 −1.5–2.9 5.8 2.0–9.6 6.2 2.3–10.1
Autologous LD flap 8.6 5.4–11.8 4.9 1.8–8.0 −1.4 −3.2–0.4 4.3 0.2–8.4 3.0 −0.6–6.6
Pedicled TRAM flap 21.2 11.0–31.4 14.1 6.8–21.4 4.6 0.4–8.8 21.1 7.2–35.0 17.5 10.7–24.2
Free TRAM flap 5.2 0.2–10.2 2.6 −2.6–7.8 −1.3 −4.6–1.9 5.1 −2.1–12.4 3.5 −1.3–8.4
DIEP/SIEA flap 9.3 5.6–13.1 5.9 2.2–9.6 1.4 −0.5–3.4 7.1 2.3–11.8 4.3 −0.8–9.3
Wald test of differ-
ence, p <0.0001 <0.0001 0.0074 0.0030 0.0001
LD, latissimus dorsi.

Table 4. Adjusted Mean Difference in BREAST-Q Scores by Type of Reconstruction, with Expander/Implant as
Reference, for Each Delayed Reconstruction Procedure Type
Breast Area Emotional Physical Sexual Satisfaction with
Appearance Well-being Well-being Well-being Outcome
Reconstruction Mean Mean Mean Mean Mean
Type Difference 95% CI Difference 95% CI Difference 95% CI Difference 95% CI Difference 95% CI
Expander/implant Reference
LD flap with implant
or expander 9.5 4.7–14.3 8.2 2.6–13.8 −1.4 −6.4–3.6 7.1 −1.9–16.1 7.9 2.1–13.6
Autologous LD flap 11.7 6.1–17.2 7.1 0.7–13.6 −0.4 −4.3–3.6 4.6 −5.2–14.4 8.0 1.0–15.0
Pedicled TRAM flap 9.4 2.6–16.2 6.2 −2.6–15.0 −1.7 −8.7–5.3 5.7 −7.8–19.1 2.8 −6.3–11.8
Free TRAM flap 19.2 12.9–25.5 10.8 3.3–18.4 −0.1 −4.9–4.6 15.3 3.6–27.1 14.9 8.1–21.8
DIEP/SIEA flap 13.8 8.2–19.5 12.1 5.5–18.7 4.0 −0.2–8.2 12.0 1.0–23.0 13.9 7.5–20.2
Wald test of differ-
ence, p <0.0001 0.0054 0.0084 0.0211 <0.0001
LD, latissimus dorsi.

Emotional Well-being (1-point difference for to endure. We therefore suggest that the results
implant-only; 5- to 16-point difference for autolo- should only be used to inform women of the out-
gous procedures) and Sexual Well-being (7-point comes they might expect to achieve with different
difference for implant-only; 13- to 28 -point differ- treatment scenarios and should be used alongside
ence for autologous procedures), whereas Physi- information about the surgical approach, compli-
cal Well-Being scores were similar in those who cation rates, and recovery time.
underwent mastectomy or any form of immediate
reconstruction. There is no agreed definition of Strengths and Limitations
what constitutes a clinically important difference Our study had a number of strengths. First,
for the BREAST-Q scales used in our study, but we used specifically developed and previously
a difference of approximately 10 points on each validated outcome measures that were distributed
scale may be taken to be equivalent to a moder- centrally at a standardized follow-up interval of
ate effect size using Norman’s standard formula 18 months, to ensure that the great majority of
of 0.5 SD.23 women had completed any adjuvant treatments
We would caution that these results should not and secondary reconstructive procedures requi
be interpreted as a prescriptive indication in favor red.2,17,22,24,25 Second, we examined outcomes for
of one set of procedures over another. We were a national population that was recruited prospec-
not able to collect data on women’s perspectives tively, with explicit written consent obtained before
on what might have led them to select one type inclusion. Third, we had excellent response rates
of reconstruction over another. Consequently, we of close to 80 percent. Fourth, we included women
have no way of taking into account their baseline with failed reconstructive procedures within the
expectations or the influence of their preferences study to minimize bias. Fifth, to minimize the risk
for the level of surgical insult they were willing of confounding, we undertook robust case-mix

1042e
Copyright © 2017 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Volume 139, Number 5 • Breast Reconstruction Patient Outcomes

adjustment within a multiple linear regression wall and cannot be positioned to extend below
model that included all factors that significantly the inframammary fold, unlike native breast tissue
affected our study outcomes. This adjustment had or a flap. An implant can usually only be placed
little effect on the outcome scores of each scale, safely under mastectomy skin if covered by an
suggesting that our procedure-specific compari- additional layer, whether muscle or, alternatively,
sons were safe and robust. an acellular dermal matrix or dermal flap or sling.
There were also some weaknesses. Although If not using a flap, the standard approach is to
the overall sample of approximately 1500 imme- place an expander under the pectoralis major and
diate reconstruction and 700 delayed recon- stretch the muscle out, as otherwise the muscle is
struction patients was large compared with other inadequate in terms of coverage and leaves the
studies in this field, the size of certain subgroups lower pole of the implant exposed. Using a sub-
(e.g., those who had a pedicled TRAM flap) was muscular implant or expander placement without
relatively small, with wide confidence intervals breast tissue to provide additional cover means
and an increased risk of sampling bias. that the final breast mound shape achieved is dif-
Next, not all eligible women were asked to ficult to control and predict. Although this may
participate in the study, and it was not possible to be less of an issue with bilateral reconstruction,
estimate the recruitment rates for specific proce- our study included only those women with unilat-
dures. It is possible that recruitment was higher eral reconstruction whose contralateral breast was
for certain procedures and that this has intro- preserved.
duced a bias to the estimated outcomes we have Finally, implants lead the body to form a
recorded. However, there is no obvious reason capsule of scar tissue around them. This capsu-
why recruitment may have been higher for cer- lar tissue contracts in a significant proportion of
tain procedures, and the diffusion of patients in patients, more commonly following radiotherapy.
different procedure groups across a very large An implant with a contracted capsule sits proud
number of hospitals makes it unlikely that vari- on the chest wall and may cause pain in addi-
able recruitment introduced a systematic error to tion to distortion. These implants may need to be
our findings. removed or exchanged with capsule management
There was also the potential for heterogeneity (excision or release) at more frequent intervals
of outcome across the large number of hospitals for those women who are worst affected, resulting
we included in the study. However, we included in a lifetime of additional procedures for some.
the identity of these organizations as a variable In contrast, using fat with or without muscle
in our regression models to adjust for any orga- and skin from the back, abdomen, buttock, or
nization-level clustering. Finally, our comparisons thigh (a flap) has a number of advantages. First,
do not include recently developed reconstructive because they consist of the patient’s own tissues,
techniques such as lipomodeling and acellular they grow and shrink with the patient, and the
dermal matrix and dermal sling procedures that contralateral breast, as their weight changes. Sec-
were not widely used during the study period. ond, as flap constituents are similar to the breast
tissue excised, they are better able to mimic the
contralateral breast’s natural shape, ptosis, and
Interpretation of Our Findings movement. Third, infection risk is much lower, as
This study is consistent with a growing body of the tissues have an intact or restored blood supply
evidence that reconstruction using patients’ own and can respond to pathogens in a normal man-
tissues is associated with better aesthetic outcomes ner. Fourth, there are no capsular issues, and the
than reconstruction using only an implant or reconstruction, if primarily successful, is definitive
expander.6–14 There are a number of explanations and lasts for life in most cases without the need for
for why autologous procedures are associated with replacement.
superior aesthetic outcomes. The principal disadvantage of flap-based
An implant or expander on its own can replace reconstruction is the need for a flap donor site
the volume that has been lost during mastectomy that will inevitably be left with a scar, and some-
but cannot fully reconstitute the breast mound times with a contour defect, muscle weakness or
with respect to its shape or position on the chest bulge, sensory changes, or another type of long-
wall. They are also static devices that will not adjust standing or permanent morbidity. Another disad-
or change automatically with a patient’s body hab- vantage is that these procedures are usually longer
itus. Implants generally do not produce natural in duration and may expose patients to a higher
ptosis of the breast, as they adhere to the chest risk of distant and systemic complications.

1043e
Copyright © 2017 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Plastic and Reconstructive Surgery • May 2017

If an implant is used in combination with a flap ACKNOWLEDGMENTS

(e.g., in the latissimus dorsi with implant group), This publication is based on data collected by or
the flap provides many of the advantages mentioned on behalf of the Healthcare Quality Improvement Part-
above, and ameliorates some of the aesthetic limi- nership, which has no responsibility or liability for the
tations of implant-only reconstruction by providing accuracy, currency, reliability, and/or correctness of
soft-tissue coverage and a more natural shape, and this publication. We have only been able to report these
allowing definitive implant placement without the findings because of the tremendous participation across
need for a temporizing expander. This is reflected National Health Service and independent hospital orga-
in the high outcomes scores associated with such nizations, the support of the professional bodies and
procedures, which in this study were comparable to patient groups involved in breast cancer care, and the
those for autologous flap reconstruction. However, engagement and participation of women with breast
the patient is left with both a degree of donor-site cancer who completed questionnaires to report on their
morbidity and the need for capsule management postoperative outcomes, and we would like to thank
and implant replacement over time. them for this. We would especially like to thank Andrea
Pusic, Stefan Cano, Anne Klassen, and their colleagues
Implications for Clinical Practice for allowing us to use the BREAST-Q Mastectomy and
Clinicians should ensure that women are pro- Breast Reconstruction modules for this national prospec-
vided with appropriate and adequate information tive study.
about all breast reconstruction options as part of
the decision-making process, including our prin-
cipal finding that autologous procedures are asso- REFERENCES
ciated with better aesthetic outcomes. 1. Jeevan R, Mennie JC, Mohanna PN, O’Donoghue JM,
Our case-mix findings suggest that although Rainsbury RM, Cromwell DA. National trends and regional
there may be some selection of reconstructive variation in immediate breast reconstruction rates. Br J Surg.
2016;103:1147–1156.
procedure type by indication, the populations 2. Pusic A, Klassen A, Scott A, Klok J, Cordeiro P, Cano S.
undergoing different procedures are broadly Development of a new patient reported outcome mea-
comparable. This suggests that many women sure for breast surgery: The BREAST-Q. Plast Reconstr Surg.
who undergo implant-based reconstruction may 2009;124:345–353.
also be suitable for flap-based reconstruction. 3. Pusic AL, Chen CM, Cano S, et al. Measuring quality of life
in cosmetic and reconstructive breast surgery: A system-
Although not all women would choose autolo- atic review of patient-reported outcomes instruments. Plast
gous procedures because of their greater com- Reconstr Surg. 2007;120:823–837; discussion 838.
plexity, longer duration, donor-site morbidity, and 4. Winters ZE, Benson JR, Pusic AL. A systematic review of
higher overall complication rates, they should be the clinical evidence to guide treatment recommendations
informed that implant-based procedures are not in breast reconstruction based on patient- reported out-
come measures and health-related quality of life. Ann Surg.
without risks, particularly implant failure,16 and 2010;252:929–942.
have now been shown to be associated with lower 5. Beesley H, Ullmer H, Holcombe C, Salmon P. How patients
patient-reported outcome scores. evaluate breast reconstruction after mastectomy, and why
their evaluation often differs from that of their clinicians.
J Plast Reconstr Aesthet Surg. 2012;65:1064–1071.
CONCLUSIONS 6. Lindegren A, Halle M, Docherty Skogh AC, Edsander-Nord
A. Postmastectomy breast reconstruction in the irradiated
Women who accessed autologous techniques breast: A comparative study of DIEP and latissimus dorsi flap
tended to report greater satisfaction with their outcome. Plast Reconstr Surg. 2012;130:10–18.
reconstruction than those who had an expander 7. Liu C, Zhuang Y, Momeni A, et al. Quality of life and patient
or implant. We recommend that clinicians act on satisfaction after microsurgical abdominal flap versus staged
expander/implant breast reconstruction: A critical study of
our findings by providing all patients with appro-
unilateral immediate breast reconstruction using patient-
priate written and verbal information regarding reported outcomes instrument BREAST-Q. Breast Cancer Res
the full range of reconstructive options, including Treat. 2014;146:117–126.
those not available at their treating hospital, and 8. Yueh JH, Slavin SA, Adesiyun T, et al. Patient satisfaction
their associated patient-reported outcomes. in postmastectomy breast reconstruction: A comparative
evaluation of DIEP, TRAM, latissimus flap, and implant tech-
David A. Cromwell, Ph.D. niques. Plast Reconstr Surg. 2010;125:1585–1595.
Clinical Effectiveness Unit 9. Hu ES, Pusic AL, Waljee JF, et al. Patient-reported aesthetic
The Royal College of Surgeons of England satisfaction with breast reconstruction during the long-term
35-43 Lincoln’s Inn Fields survivorship period. Plast Reconstr Surg. 2009;124:1–8.
London WC2A 3PE, United Kingdom 10. Tønseth KA, Hokland BM, Tindholdt TT, Abyholm FE,
dcromwell@rcseng.ac.uk Stavem K. Quality of life, patient satisfaction and cosmetic

1044e
Copyright © 2017 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Volume 139, Number 5 • Breast Reconstruction Patient Outcomes

outcome after breast reconstruction using DIEP flap or 17. Cohen WA, Mundy LR, Ballard TN, et al. The BREAST-Q in
expandable breast implant. J Plast Reconstr Aesthet Surg. surgical research: A review of the literature 2009-2015. J Plast
2008;61:1188–1194. Reconstr Aesthet Surg. 2016;69:149–162.
11. Eltahir Y, Werners LL, Dreise MM, Zeijlmans van 18. Noble M, McLennan D, Wilkinson K, et al. The English
Emmichoven IA, Werker PM, de Bock GH. Which breast is Indices of Deprivation 2007: Summary. London: Her Majesty’s
the best? Successful autologous or alloplastic breast recon- Stationary Office; 2008.
struction: Patient-reported quality-of-life outcomes. Plast 19. Dripps RD. New classification of physical status. Anesthesiology
Reconstr Surg. 2015;135:43–50. 1963;24:111.
12. Alderman AK, Wilkins EG, Lowery JC, Kim M, Davis JA. 20. Oken MM, Creech RH, Tormey DC, et al. Toxicity and
Determinants of patient satisfaction in postmastectomy response criteria of the Eastern Cooperative Oncology
breast reconstruction. Plast Reconstr Surg. 2000;106:769–776. Group. Am J Clin Oncol. 1982;5:649–655.
13. Weichman KE, Hamill JB, Kim HM, Chen X, Wilkins EG, 21. Haybittle JL, Blamey RW, Elston CW, et al. A prog-
Pusic AL. Understanding the recovery phase of breast recon- nostic index in primary breast cancer. Br J Cancer
structions: Patient-reported outcomes correlated to the type 1982;45:361–366.
and timing of reconstruction. J Plast Reconstr Aesthet Surg. 22. Cano SJ, Klassen AF, Scott A, Alderman A, Pusic AL.
2015;68:1370–1378. Interpreting clinical differences in BREAST-Q scores:
14. Atisha DM, Cox CE, Rushing CN, Samsa GP, Pusic AL, Minimal important difference. Plast Reconstr Surg.
Abernethy AP. A national snapshot of satisfaction with breast 2014;134:173e–175e.
cancer procedures. Ann Surg Oncol. 2015;22:361–369. 23. Norman GR, Sloan JA, Wyrwich KW. Interpretation of
15. Spear SL, Newman MK, Bedford MS, Schwartz KA, Cohen changes in health-related quality of life: The remark-
M, Schwartz JS. A retrospective analysis of outcomes using able universality of half a standard deviation. Med Care
three common methods for immediate breast reconstruc- 2003;41:582–592.
tion. Plast Reconstr Surg. 2008;122:340–347. 24. Cano SJ, Klassen AF, Scott AM, Cordeiro PG, Pusic AL. The
16. Jeevan R, Cromwell DA, Browne JP, et al. Findings of a BREAST-Q: Further validation in independent clinical sam-
national comparative audit of mastectomy and breast recon- ples. Plast Reconstr Surg. 2012;129:293–302.
struction surgery in England. J Plast Reconstr Aesthet Surg. 25. Cano SJ, Klassen AF, Scott AM, Pusic AL. A closer look at the
2014;67:1333–1344. BREAST-Q. Clin Plast Surg. 2013;40:287–296.

1045e
Copyright © 2017 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.