Nationwide population-based study of the impact of immediate breast reconstruction after mastectomy on the timing of adjuvant chemotherapy.

BACKGROUND: Initiation of adjuvant chemotherapy within 6-12 weeks after mastectomy is recommended by guidelines. The aim of this population-based study was to investigate whether immediate breast reconstruction (IBR) after mastectomy reduces the likelihood of timely initiation of adjuvant chemotherapy.
METHODS: All patients with breast cancer who had undergone mastectomy and adjuvant chemotherapy between 2012 and 2016 in the Netherlands were identified. Time from surgery to adjuvant chemotherapy was categorized as within 6 weeks or after more than 6 weeks, within 9 weeks or after more than 9 weeks, and within 12 weeks or after more than 12 weeks. The impact of IBR on the initiation of adjuvant chemotherapy for these three scenarios was estimated using propensity score matching to adjust for treatment by indication bias.
RESULTS: A total of 6300 patients had undergone primary mastectomy and adjuvant chemotherapy, of whom 1700 (27·0 per cent) had received IBR. Multivariable analysis revealed that IBR reduced the likelihood of receiving adjuvant chemotherapy within 6 weeks (odds ratio (OR) 0·76, 95 per cent c.i. 0·66 to 0·87) and 9 weeks (0·69, 0·54 to 0·87), but not within 12 weeks (OR 0·75, 0·48 to 1·17). Following propensity score matching, IBR only reduced the likelihood of receiving adjuvant chemotherapy within 6 weeks (OR 0·95, 0·90 to 0·99), but not within 9 weeks (OR 0·97, 0·95 to 1·00) or 12 weeks (OR 1·00, 0·99 to 1·01).
CONCLUSION: Postmastectomy IBR marginally reduced the likelihood of receiving adjuvant chemotherapy within 6 weeks, but not within 9 or 12 weeks. Thus, IBR is not contraindicated in patients who need adjuvant chemotherapy after mastectomy.

Introduction

Breast cancer is the most commonly diagnosed malignant cancer among women1. Despite advancements in diagnostics and systemic treatment, up to one‐third of patients with breast cancer undergo mastectomy as the first surgical treatment to achieve local control2. Adjuvant systemic treatment, including chemotherapy, reduces the risk of distant recurrence and breast cancer mortality3. In the Netherlands, 6 weeks is the maximum time limit aimed for between surgery and initiation of adjuvant chemotherapy, as recommended by the European Society for Medical Oncology4 (ESMO) and the Netherlands Society for Plastic Surgery5, 6.

Several studies7, 8, 9, 10, 11, 12 have reported that delayed initiation of adjuvant chemotherapy is associated with lower overall and recurrence‐free survival. The recommended acceptable maximum delay, however, varies from 7 to 12 weeks. There still is no international consensus on the definition of an unacceptable delay, but all guidelines advocate that initiation of adjuvant chemotherapy should not be delayed unnecessarily, as this may have a negative impact on survival, specifically in patients at higher risk of recurrence9, 10, 12.

The addition of immediate breast reconstruction (IBR) to mastectomy could result in preoperative delay owing to more complex logistic coordination of the operation. After surgery, a delay could be the result of longer recovery, as IBR may increase the risk of postoperative complications, even though reports on the risk of adverse events are conflicting13, 14, 15, 16.

In the past decade, an increasing number of women have undergone IBR after mastectomy2, 17, 18. IBR is generally associated with good aesthetic results and less negative psychological impact on the patient, as it involves fewer operations and hospital admissions compared with breast reconstruction at a later time19, 20, 21. Owing to the lack of consensus on timing of adjuvant chemotherapy, physicians remain cautious in recommending IBR when adjuvant chemotherapy is part of the preoperative treatment plan22.

Most previous studies16, 23 on the possible delaying impact of postmastectomy IBR have been single‐centre studies with weak methodology and no adjustment for treatment by indication bias. A systematic review24 from 2015 concluded that IBR does not delay time from surgery to adjuvant chemotherapy to a clinically relevant extent, although the included studies showed strongly contradictory results. Moreover, a cut‐off point of 12 weeks to initiation of adjuvant treatment was used, whereas current European guidelines4 recommend 6 weeks. Furthermore, it seems likely that there may be an underlying reason why some patients have IBR and others do not, giving rise to treatment by indication bias when comparing the outcomes of these two groups.

The aim of the present nationwide population‐based study was to investigate the extent to which postmastectomy IBR reduces the likelihood of timely initiation of adjuvant chemotherapy compared with mastectomy alone, while also adjusting for confounding by indication.

Methods

Prospectively collected data from the NABON Breast Cancer Audit (NBCA) database were used. The NBCA was started in 2011 and is an initiative from the National Breast Cancer Organization Netherlands (NABON), the Netherlands Comprehensive Cancer Organization and the Dutch Institute for Clinical Auditing. The NBCA collects anonymized data on clinicopathological characteristics, diagnostics and treatment modalities in a database from all hospitals in the Netherlands. It includes all patients diagnosed with ductal carcinoma in situ (DCIS) or invasive breast cancer treated surgically since 2012. The NBCA aims to monitor the quality of breast cancer care and to provide feedback to participating hospitals to stimulate and facilitate quality improvement25. No formal consent is required for this type of study from an ethics committee in the Netherlands according to Central Committee on Research involving Human Subjects.

Patient population

All women diagnosed with invasive breast cancer between 2012 and 2016 who had undergone primary mastectomy with or without IBR followed by adjuvant chemotherapy were identified from the NBCA database. IBR was defined as a reconstruction performed by a plastic surgeon on the same day as the mastectomy. Women who had received systemic neoadjuvant treatment, had undergone lumpectomy as initial surgery or had a re‐excision were excluded from the analysis. Patients who had received another adjuvant therapy before the initiation of adjuvant chemotherapy, and those with a missing date of operation or adjuvant chemotherapy were also excluded.

Outcomes

The primary outcome was whether the patient received adjuvant chemotherapy within a specific time interval after surgery. Time to adjuvant chemotherapy was analysed with three different cut‐off values: within 6 weeks or after more than 6 weeks, within 9 weeks or after more than 9 weeks, and within 12 weeks or after more than 12 weeks. These cut‐offs were chosen based on the currently recommended starting point according to Dutch and ESMO guidelines, and on previous literature demonstrating that a clinical impact is found when adjuvant chemotherapy is started later than 7–12 weeks, indicating the importance of initiating adjuvant chemotherapy at least within this time period4, 7, 8, 9, 10, 11, 12.

Confounders

Potential confounders included in analyses were year of diagnosis, age, WHO performance status26, presence of DCIS, histological type, receptor status, tumour stage according to the seventh edition of AJCC27, sentinel node biopsy, axillary lymph node dissection (ALND), hospital transfer between site for surgery and that for adjuvant chemotherapy, and annual number of patients operated on for breast cancer at the hospital (hospital volume). Data regarding reconstruction at a later time, rather than IBR, are not registered in the NBCA and could therefore not be included.

Statistical analysis

Statistical differences for all possible confounders between women who had mastectomy alone and those who had mastectomy plus IBR were determined using χ2 tests. All tests were two‐sided, and P < 0·050 was considered statistically significant. Multivariable logistic regression analysis was used to determine the likelihood that women who had undergone IBR received adjuvant chemotherapy within 6, 9 and 12 weeks, when adjusted for the confounders. There may, however, be an underlying reason why patients have IBR, so that not all women are equally likely to receive IBR, for example because of a different type of tumour or age of the patient, introducing a treatment by indication bias. Thus, propensity score matching (PSM) was performed, including all available patient and tumour characteristics to adjust for treatment by indication bias. Use of PSM ensures that patients from both cohorts are matched and have the same likelihood of receiving IBR, given certain patient and tumour characteristics. For each pair, one patient did and one did not undergo IBR; this is essential to estimate the true treatment effect on an outcome in observational studies28, 29. Statistical analyses were performed with SPSS® version 24 (IBM, Armonk, New York, USA).

Results

In the selected time interval, 6300 women were diagnosed with invasive breast cancer and met the eligibility criteria. Of these, 4600 patients (73·0 per cent) underwent mastectomy alone and 1700 patients (27·0 per cent) had postmastectomy IBR. Of the women who had IBR, 91·2 per cent had received an implant‐based reconstruction (including tissue expanders).

The proportion of women who had postmastectomy IBR decreased with patient age and increased over time (Fig. 1). Patients who underwent IBR were younger at diagnosis, more often had a WHO status of 0, or were diagnosed with no special type of histology, DCIS component and tumour stage I than women who had mastectomy alone (Table 1). There was no difference in receptor status or differentiation grade between the two groups. Of women who had postmastectomy IBR, the proportions that underwent sentinel node biopsy, transferred hospital between surgery and adjuvant chemotherapy, or were treated in a hospital with surgical volume exceeding 250 patients annually were also higher compared with those of women who had mastectomy alone. However, the proportion that had ALND was lower in women who underwent postmastectomy IBR (Table 1).

Figure 1

Proportion of patients having immediate breast reconstruction in relation to year of diagnosis and age Immediate breast reconstruction (IBR) in relation to

Table 1

Baseline characteristics of patients who had mastectomy alone or immediate breast reconstruction after mastectomy and received adjuvant chemotherapy

	Mastectomy alone (n = 4600)	IBR after mastectomy (n = 1700)	P *
Year of diagnosis			< 0·001
2012	1282 (27·9)	290 (17·1)
2013	1113 (23·5)	365 (21·5)
2014	987 (21·5)	378 (22·2)
2015	690 (15·0)	411 (24·2)
2016	528 (11·5)	256 (15·1)
Age (years)			< 0·001
< 40	304 (6·6)	295 (17·4)
40–49	1081 (23·5)	578 (34·0)
50–59	1506 (32·7)	578 (34·0)
60–69	1409 (30·6)	233 (13·7)
≥ 70	300 (6·5)	16 (0·9)
WHO performance status			0·001
0	4126 (89·7)	1572 (92·5)
1	450 (9·8)	116 (6·8)
≥ 2	24 (0·5)	12 (0·7)
Histology			< 0·001
No special type	3580 (77·8)	1414 (83·2)
Lobular	731 (15·9)	168 (9·9)
Both/other	289 (6·3)	118 (6·9)
DCIS component			< 0·001
No	2241 (48·7)	623 (36·6)
Yes	2359 (51·3)	1077 (63·4)
Receptor status			0·071
Triple‐negative	695 (15·1)	223 (13·1)
HER2‐neu+	1053 (22·9)	405 (23·8)
HR+ and HER2−	2727 (59·3)	1038 (61·1)
Unknown	125 (2·7)	34 (2·0)
Differentiation grade			0·987
Well	431 (9·4)	161 (9·5)
Moderate	2136 (46·4)	791 (46·5)
Poor	2033 (44·2)	748 (44·0)
Tumour stage			< 0·001
I	1036 (22·5)	735 (43·2)
IIa	1542 (33·5)	632 (37·2)
IIb	856 (18·6)	200 (11·8)
III	1128 (24·5)	128 (7·5)
IV	38 (0·8)	5 (0·3)
Sentinel node biopsy			< 0·001
No	1439 (31·3)	131 (7·7)
Yes	3161 (68·7)	1569 (92·3)
ALND			< 0·001
No	2303 (50·1)	1265 (74·4)
Yes	2297 (49·9)	435 (25·6)
Hospital transfer			0·030
No	4466 (97·1)	1632 (96·0)
Yes	134 (2·9)	68 (4·0)
Hospital volume of surgery(no. of patients)			< 0·001
1–99	223 (4·8)	29 (1·7)
100–149	1036 (22·5)	263 (15·5)
150–199	978 (21·3)	253 (14·9)
200–249	478 (10·4)	236 (13·9)
≥ 250	1885 (41·0)	919 (54·1)

Values in parentheses are percentages. IBR, immediate breast reconstruction; DCIS, ductal carcinoma in situ; HR+, hormone receptor‐positive; ALND, axillary lymph node dissection.

χ2 test.

Time to adjuvant chemotherapy

The median (i.q.r.) time from surgery to adjuvant chemotherapy in women who had postmastectomy IBR was 36 (29–47) days, compared with 34 (28–44) days in those who had mastectomy alone (Table 2). Adjuvant chemotherapy was initiated within 6 weeks in more than two‐thirds of patients, and the vast majority received adjuvant chemotherapy within 9 and 12 weeks. The total proportion of patients who received adjuvant chemotherapy within 6, 9 and 12 weeks did not differ over time (2012–2016: P = 0·282, P = 0·128 and P = 0·052 respectively) (Fig. 2).

Table 2

Time from surgery to adjuvant chemotherapy, and proportion of patients receiving adjuvant chemotherapy within 6, 9 and 12 weeks

	Mastectomy alone (n = 4600)	IBR after mastectomy (n = 1700)
Time from surgery to adjuvantchemotherapy (days) *	34 (28–44)	36 (29–47)
No. of patients receiving adjuvantchemotherapy
Within 6 weeks	3297 (71·7)	1145 (67·4)
Within 9 weeks	4304 (93·6)	1564 (92·0)
Within 12 weeks	4509 (98·0)	1669 (98·2)

Values in parentheses are percentages unless indicated otherwise;

values are median (i.q.r.). IBR, immediate breast reconstruction.

Figure 2

Proportion of women receiving adjuvant chemotherapy within 6, 9 and 12 weeks in relation to year of diagnosis

Unmatched multivariable analyses

Multivariable analysis revealed that patients who had undergone IBR were less likely than those having mastectomy alone to receive adjuvant chemotherapy within 6 weeks (odds ratio (OR) 0·76, 95 per cent c.i. 0·66 to 0·87; P < 0·001) or 9 weeks (OR 0·69, 0·54 to 0·87; P = 0·002) of surgery (Table 3). However, IBR had no association with receiving adjuvant chemotherapy within 12 weeks (OR 0·75, 0·48 to 1·17; P = 0·205).

Table 3

Univariable and multivariable analyses without propensity score matching of characteristics associated with time to adjuvant chemotherapy within 6, 9 and 12 weeks

		Time to adjuvant chemotherapy
		≤ 6 weeks		≤ 9 weeks		≤ 12 weeks
	No. of patients (n = 6300)*	OR (univariable)	OR (multivariable)	OR (univariable)	OR (multivariable)	OR (univariable)	OR (multivariable)
IBR after mastectomy
No	4600 (73·0)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)
Yes	1700 (27·0)	0·82 (0·72, 0·92)	0·76 (0·66, 0·87)	0·79 (0·64, 0·98)	0·69 (0·54, 0·87)	1·09 (0·72, 1·64)	0·75 (0·48, 1·17)
Year of diagnosis
2012	1572 (25·0)	1·00 (reference)	–	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)
2013	1478 (23·5)	1·03 (0·88, 1·21)	–	0·96 (0·73, 1·26)	0·95 (0·72, 1·25)	1·28 (0·80, 2·05)	1·30 (0·81, 2·08)
2014	1365 (21·7)	1·12 (0·95, 1·31)	–	1·11 (0·83, 1·48)	1·05 (0·78, 1·42)	1·53 (0·92, 2·55)	1·50 (0·90, 2·50)
2015	1101 (17·5)	0·99 (0·83, 1·17)	–	1·43 (1·03, 1·99)	1·47 (1·04, 2·07)	2·49 (1·31, 4·75)	2·44 (1·26, 4·70)
2016	784 (12·4)	0·91 (0·76, 1·09)	–	0·94 (0·68, 1·31)	0·85 (0·60, 1·20)	1·63 (0·87, 3·05)	1·52 (0·80, 2·89)
Age (years)
< 40	599 (9·5)	1·13 (0·92, 1·39)	1·17 (0·94, 1·46)	1·17 (0·79, 1·72)	1·17 (0·78, 1·75)	1·23 (0·56, 2·66)	1·28 (0·59, 2·79)
40–49	1659 (26·3)	1·18 (1·02, 1·37)	1·20 (1·03, 1·40)	1·24 (0·94, 1·63)	1·21 (0·92, 1·60)	0·93 (0·57, 1·54)	0·94 (0·57, 1·55)
50–59	2084 (33·1)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)
60–69	1642 (26·1)	0·78 (0·68, 0·93)	0·68 (0·59, 0·79)	0·72 (0·56, 0·91)	0·64 (0·49, 0·82)	0·60 (0·38, 0·95)	0·57 (0·36, 0·89)
≥ 70	316 (5·0)	0·71 (0·55, 0·91)	0·51 (0·39, 0·67)	0·82 (0·53, 1·28)	0·62 (0·39, 0·99)	0·73 (0·32, 1·67)	0·68 (0·30, 1·56)
WHO performance status
0	5698 (90·4)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	–
1	566 (9·0)	0·62 (0·52, 0·74)	0·62 (0·51, 0·75)	0·63 (0·47, 0·85)	0·63 (0·46, 0·85)	0·75 (0·43, 1·31)	–
≥ 2	36 (0·6)	0·44 (0·23, 0·86)	0·51 (0·25, 1·02)	0·35 (0·14, 0·84)	0·39 (0·15, 0·98)	0·32 (0·08, 1·36)	–
Histology
No special type	4994 (79·3)	1·00 (reference)	–	1·00 (reference)	–	1·00 (reference)	–
Lobular	899 (14·3)	0·96 (0·82, 1·12)	–	1·04 (0·78, 1·38)	–	1·67 (0·89, 3·12)	–
Both/other	407 (6·5)	0·86 (0·69, 1·06)	–	0·80 (0·55, 1·15)	–	0·82 (0·42, 1·58)	–
DCIS component
No	2864 (45·5)	1·00 (reference)	–	1·00 (reference)	–	1·00 (reference)	–
Yes	3436 (54·5)	0·99 (0·89, 1·11)	–	0·90 (0·74, 1·10)	–	0·89 (0·62, 1·28)	–
Receptor status
Triple‐negative	918 (14·6)	1·34 (1·14, 1·58)	1·12 (1·03, 1·22)	1·33 (0·99, 1·80)	0·96 (0·69, 1·35)	0·79 (0·49, 1·29)	–
HER‐2+	1458 (23·1)	1·34 (1·17, 1·53)	1·17 (1·09, 1·26)	1·43 (1·11, 1·85)	1·19 (0·91, 1·57)	1·12 (0·71, 1·77)	–
HR+/HER2−	3765 (59·8)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	–
Unknown	159 (2·5)	1·50 (1·03, 2·17)	1·94 (1·70, 2·22)	1·39 (0·70, 2·74)	1·51 (0·75, 3·06)	1·01 (0·32, 3·25)	–
Differentiation grade
Well	592 (9·4)	0·70 (0·58, 0·84)	0·90 (0·73, 1·11)	0·55 (0·40, 0·75)	0·68 (0·48, 0·96)	0·61 (0·35, 1·05)	–
Moderate	2927 (46·5)	0·83 (0·74, 0·93)	0·94 (0·85, 1·11)	0·71 (0·57, 0·88)	0·81 (0·64, 1·03)	1·05 (0·72, 1·55)	–
Poor	2781 (44·1)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	–
Tumour stage
I	1771 (28·1)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	–
IIa	2174 (34·5)	1·08 (0·94, 1·24)	1·44 (1·24, 1·68)	1·12 (0·87, 1·45)	1·51 (1·14, 2·00)	1·38 (0·87, 2·20)	–
IIb	1056 (16·8)	0·72 (0·61, 0·84)	1·30 (1·06, 1·60)	0·73 (0·55, 0·97)	1·34 (0·94, 1·90)	0·99 (0·58, 1,66)	–
III	1256 (19·9)	1·11 (0·94, 1·30)	1·72 (1·37, 2·15)	0·90 (0·67, 1·19)	1·43 (0·98, 2·09)	1·03 (0·63, 1·70)	–
IV	43 (0·7)	0·52 (0·28, 0·95)	0·65 (0·34, 1·25)	2·97 (0·41, 21·78)	3·76 (0·50, 28·18)	–	–
Sentinel node biopsy
No	1570 (24·9)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	–
Yes	4730 (75·1)	0·51 (0·44, 0·58)	0·23 (0·19, 0·27)	0·59 (0·46, 0·77)	0·33 (0·24, 0·45)	0·85 (0·56, 1·31)	–
ALND
No	3568 (56·6)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)
Yes	2732 (43·4)	0·57 (0·51, 0·63)	0·23 (0·19, 0·27)	0·53 (0·44, 0·65)	0·30 (0·23, 0·39)	0·56 (0·39, 0·81)	0·58 (0·40, 0·85)
Hospital transfer†
Same hospital	6098 (96·8)	1·00 (reference)	1·00 (reference)	1·00 (reference)	–	1·00 (reference)	–
Different hospital	202 (3·2)	0·55 (0·42, 0·73)	0·48 (0·36, 0·66)	0·75 (0·45, 1·22)	–	0·98 (0·36, 2·67)	–
Hospital volume of surgery(no. of patients)
1–99	252 (4·0)	0·91 (0·67, 1·23)	0·94 (0·68, 1·30)	1·37 (0·70, 2·70)	1·40 (0·70, 2·79)	1·37 (0·40, 4·65)	–
100–149	1299 (20·6)	0·88 (0·74, 1·04)	0·87 (0·72, 1·04)	0·70 (0·51, 0·97)	0·71 (0·51, 0·99)	0·70 (0·40, 1·24)	–
150–199	1231 (19·5)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	1·00 (reference)	–
200–249	714 (11·3)	0·96 (0·78, 1·20)	0·61 (0·76, 1·18)	0·68 (0·47, 0·98)	0·69 (0·47, 1·00)	0·60 (0·32, 1·14)	–
≥ 250	2804 (44·5)	0·76 (0·66, 0·89)	0·75 (0·66, 0·87)	0·72 (0·54, 0·96)	0·76 (0·57, 1·02)	0·91 (0·54, 1·54)	–

Values in parentheses are 95 per cent confidence intervals unless indicated otherwise;

values are number (per cent). †Between surgery and adjuvant chemotherapy. IBR, immediate breast reconstruction; DCIS, ductal carcinoma in situ; HR+, hormone receptor‐positive; ALND, axillary lymph node dissection.

Although not the focus of this study, analyses of predictive confounders demonstrated that, amongst other factors, patients who had a sentinel node biopsy or ALND were less likely to receive adjuvant chemotherapy within 6 and 9 weeks, as well as within 12 weeks for ALND (Table 3).

Matched comparison of the two groups

Following PSM of patients with an equal likelihood of receiving IBR based on patient and tumour characteristics, women who had IBR were still less likely to receive adjuvant chemotherapy within 6 weeks (OR 0·95, 95 per cent c.i. 0·90 to 0·99; P = 0·035), but not within 9 weeks (OR 0·97, 0·95 to 1·00; P = 0·050) or 12 weeks (OR 1·00, 0·99 to 1·01; P = 0·894).

Discussion

This large population‐based study, analysing patients from all hospitals treating breast cancer in the Netherlands, found that, compared with mastectomy alone, IBR after mastectomy reduced the likelihood of receiving adjuvant chemotherapy within 6 weeks of surgery, as recommended by Dutch6 and European4, 5 guidelines, but not within 9 or 12 weeks. This suggests that postmastectomy IBR is not necessarily contraindicated in patients who need adjuvant chemotherapy, because in general IBR does not delay its initiation to a clinically relevant extent.

Previous studies on the impact of IBR on time to adjuvant chemotherapy reported a large variation in time to adjuvant chemotherapy, ranging from 21 to 80 days for those who had mastectomy alone and from 31 to 97 days for patients who received IBR30, 31, 32, 33, 34, with reported differences between these cohorts of 14–27 days24. However, this large variation may have been the result of the small single‐centre studies, weak methodology and biases, such as the lack of adjusting for treatment by indication bias.

The findings of the present study are not in line with the recently published results from a large multicentre study of Jabo and colleagues35 in the USA, which suggested that IBR delays time from diagnosis to treatment but not from surgery to adjuvant chemotherapy. This discrepancy may be explained by differences in the statistical approach, as these authors used time as a continuous value, compared with a categorical value in the present study. Moreover, Jabo and co‐workers35 compared time from surgery to adjuvant chemotherapy with non‐parametric tests without adjusting for confounders, because the latter was not the main focus of their study. It is noteworthy that their reported time from surgery to adjuvant chemotherapy was considerably longer than that found in the present study, both for patients who had mastectomy alone (40 versus 34 days respectively) and those who underwent IBR (42 versus 36 days)35.

The present study suggests that patients who had sentinel node biopsy or ALND were less likely to receive adjuvant chemotherapy within the predefined cut‐off points, confirming the previously reported delaying impact of ALND34. In the present study, postoperative complications may have occurred more frequently in patients who underwent ALND combined with postmastectomy IBR, and thereby potentially could have delayed chemotherapy36, 37, 38. Postoperative complications, such as axillary seroma, are common after mastectomy combined with ALND. The present study suggests that the associated risk of postoperative complications after sentinel node biopsy and ALND may increase the likelihood of delay. The risk of seroma formation can be reduced by minimizing dead space through quilting sutures or an axillary drain39. Complications, and strategies to prevent their occurrence, are not collected in the NBCA database and could therefore not be studied as a potential explanatory factor.

The present study has shown that patients diagnosed with triple‐negative breast cancer, human epidermal growth factor receptor 2‐positive breast cancer and higher stage disease were more likely to receive adjuvant chemotherapy within 6 weeks. It is reassuring that these tumour characteristics were predictive of timely initiation of adjuvant chemotherapy, as previous studies7, 10 have shown that delay is of particular relevance in women with these more aggressive types of cancer.

It was expected that the impact of IBR on time to adjuvant chemotherapy would change after adjusting for treatment by indication bias, as the present results and a previous Dutch study40 both showed that patients undergoing IBR differ in many characteristics from those undergoing mastectomy alone.

The majority of patients in the present study underwent a two‐stage implant IBR with a tissue expander. This type of IBR is the most common approach in patients eligible for postoperative radiotherapy in most industrialized countries41. Despite autologous reconstructions being used increasingly in the last decade18, the proportions of different types of IBR were comparable between the predefined cut‐off points (data not shown). Nonetheless, the number of women who had IBR using autologous tissue with or without a prosthesis was low (less than 8 per cent), reflecting practice in the past. Therefore, a future study with more patients receiving IBR using autologous tissue could investigate whether this will affect the results.

Patients who changed hospital after surgery were less likely to receive adjuvant chemotherapy within 6 weeks, but not within 9 or 12 weeks. Although this concerned only 3·2 per cent of all patients, the association corroborates the theory that hospital transfer delays treatment, as shown by previous studies34, 42, 43.

The present results are inconclusive regarding the association between hospital volume and time to adjuvant chemotherapy. On the one hand, higher volume reduced the likelihood of receiving adjuvant chemotherapy within 6 weeks, but on the other hand, lower volume reduced the likelihood of receiving adjuvant chemotherapy within 9 weeks. A recent study by Schreuder and co‐workers44 demonstrated that hospital volume only partly explains the use of IBR in the Netherlands. Presumably, other hospital related factors such as theatre availability or number of medical specialists have more impact on time to adjuvant chemotherapy after IBR than just hospital volume.

The number of patients aged 70 years or above seems lower in the present study than in previous studies. This might be explained by the fact that adjuvant chemotherapy is used less frequently in these older women in the Netherlands45. Furthermore, postmastectomy IBR is used less frequently in this patient group in the Netherlands40.

There were several limitations to the present study. First, it was observational, using PSM to adjust for confounding as best as possible. However, matching may be improved by adding other factors potentially associated with delay of adjuvant chemotherapy or the type of surgery (such as radiotherapy, BMI, travel distance). Unfortunately, it was not possible to include these factors as these are not registered in the NBCA database. Insurance coverage was probably not important in the present study, in contrast to studies from the USA, because all Dutch patients are obliged to have basic insurance coverage, providing equal access to breast cancer treatment and breast reconstruction. Second, treatment delay or choice for a specific type of surgery can also be the result of patient preference, such as seeking a second opinion or personal scheduling limitations. Third, this study focused on the time between surgery and initiation of adjuvant chemotherapy, and was therefore not able to assess the potential delaying impact of IBR in the preoperative phase owing to organizational factors such as planning.

The results of the present study in a population‐based setting, which were adjusted for confounding and treatment by indication bias, add to the evidence in current literature that IBR is not contraindicated in patients who require a mastectomy and adjuvant chemotherapy, because it does not generally delay time to adjuvant chemotherapy to a clinically relevant extent.

Disclosure

The authors declare no conflict of interest.