mastectomy and use of drain

European Journal of Oncology Nursing 36 (2018) 112–118
Contents lists available at ScienceDirect
European Journal of Oncology Nursing
journal homepage: www.elsevier.com/locate/ejon
Early drain removal improves quality of life and clinical outcomes in
patients with breast cancer – Results from a randomised controlled trial
T
Hanne Vosa,b,∗, Ann Smeetsa,b, Patrick Nevenb,c, Annouschka Laenend, Lienke Vandezandeb,e,
Ines Nevelsteena,b
a
Surgical Oncology, Department of Oncology, KU Leuven, University Hospitals Leuven, Belgium
Multidisciplinary Breast Centre, University Hospitals Leuven, Belgium
c
Gynaecological Oncology, Department of Oncology, KU Leuven, University Hospitals Leuven, Belgium
d
Biostatistics and Statistical Bioinformatics Centre, KU Leuven, Belgium
e
Department of Public Health and Primary Care, KU Leuven, University Hospitals Leuven, Belgium
b
A R T I C LE I N FO
A B S T R A C T
Keywords:
Suction drain
Breast cancer
Early drain removal
Quality of life
Postoperative complications
Randomised controlled trial
Purpose: The use of suction drains after breast cancer surgery (BCS) is common practice. However, the optimal
time to remove drains is not clear yet and limited research has been conducted so far to assess the impact of their
use on patient comfort. The goal of this study was to investigate the effect of early drain removal after BCS on
quality of life (QoL) and clinical outcome.
Method: A randomised controlled trial was conducted in 99 patients scheduled for BCS including placement of
suction drains. Patients were randomised into either group 1: drains removed output-based, i.e., flow less than
30 ml/day or group 2: drains removed at hospital discharge, i.e., 4–5 days after surgery. A questionnaire on QoL
was completed by the patients both pre- and postoperatively.
Results: Early drain removal was associated with a significant improvement in QoL. Additionally, total duration
of home care nursing was considerably lower in the early-removal group (19 versus 1 day on average). No
differences were observed in wound healing or the rate of wound infections, the latter being slightly lower in the
early-removal group (13% versus 6%). Total volumes of fluid drained and/or aspirated were significantly lower
in the early-removal group (median 1745 ml versus 752 ml), but more aspirations were needed (median 1 versus
3). The new policy of early drain removal was preferred by 94% of the patients in the early removal group.
Conclusions: Early removal of suction drains improves QoL and has no negative effect on clinical outcomes after
BCS.
1. Introduction
Placement of suction drains after mastectomy or axillary lymph
node dissection (ALND) has become common practice in breast cancer
surgery (BCS) to prevent seroma formation, ever since their first introduction in 1947 (Murphey, 1947). Nevertheless, placement of suction drainage has some distinct disadvantages: skin bacteria can cause
infection by retrograde entry through the drain, or the drain itself can
cause patient discomfort and a need for daily home nursing (Andeweg
et al., 2011). Likewise, drain removal policies across breast cancer (BC)
centres vary widely. Only limited prospective research has been performed to date on the impact of wound drains on postoperative quality
of life (QoL) in all its dimensions for patients with BC.
Multiple studies have investigated the safety of early drain removal
∗
based on several clinical endpoints (Table 1). No significant difference
in the incidence of seroma formation was found by Ackroyd and Reed,
1997 in a study with patients who received BCS with axillary clearance.
These patients were randomised into two groups: axillary drains were
removed after five days in the first group, while a volume-based criterion of less than 30 ml/day was used for removal in the second group.
Similarly, Okada et al. (2015) did not find a significant difference in the
incidence of seroma formation between the study and control group.
Nonetheless, the number of outpatient visits for seroma was significantly higher in the study group without significant difference in the
number of aspirations. These results were obtained in a Japanese cohort
where drain removal upon volume less than 50 ml/day or at the latest
five days postoperative was compared with conventional drain removal. Unfortunately, conventional drain removal was not specified,
Corresponding author. Oncologische Heelkunde, UZ Leuven, Herestraat, 49, 3000, Leuven, Belgium.
E-mail address: [email protected] (H. Vos).
https://doi.org/10.1016/j.ejon.2018.08.007
Received 23 April 2018; Received in revised form 26 July 2018; Accepted 17 August 2018
1462-3889/ © 2018 Elsevier Ltd. All rights reserved.
European Journal of Oncology Nursing 36 (2018) 112–118
H. Vos et al.
Table 1
Literature overview ranked by year.
Author
Ackroyd
Gupta
Dalberg
Okada
Taylor
Year
1997
2001
2004
2013
2013
Type of study
Pro
Pro
Multicentre Pro
Retro
Pro
Country
UK
India
Sweden
Japan
UK
Power calculation
Yes
No
Yes
No
No
Sample size
120
121
198
214
596
Surgery
Axillary dissection
Mastectomy BCT
Mastectomy
Mastectomy
Mastectomy BCT
Mastectomy BCT
Level I-III
ALND
Level I-II
Level I-II
ALND
Control group
Study group
QoL
Drain policy
N
Drain policy
N
< 30 ml/day
8 days
< 40 ml/day
< 50 ml/day or 5 days
< 50 ml/day or 7 days
61
57
99
76
263
5 days
5 days
1 day
Conv.
No Drain
59
64
99
138
335
Yes
No
Yes
No
No
Pro = prospective, Retro = retrospective, BCT = breast conservative therapy, ALND = axillary lymph node dissection, SLNB = sentinel lymph node biopsy,
N = number of patients, Conv = conventional and QoL = Quality of life.
Fig. 1. Trial flow.
The overall quality of the combined dataset was poor. Early drain removal was defined as removal after one to five days postoperatively,
late as removal after six to eight days or based on daily output. The
authors only mention that a higher total drainage prior to drain removal in the early drain removal group predicted subsequent seroma
formation. No significant difference in infection rate was reported between early and late removal. The length of hospital stay was significantly lower in the early-removal group. In conclusion, no optimal
timing of drain removal could be determined, a high total volume of
drainage prior to drain removal predicted seroma formation but no
exact volume was stated (Kelley et al., 2012).
Unlike the impact of early drain removal on clinical variables, the
research on the effect of drains on patient comfort and QoL is limited.
According to the authors' knowledge, only Ackroyd and Reed, 1997 and
Dalberg et al. (2004) have assessed QoL in a quantitative way. The
patient questionnaire conducted by Ackroyd and Reed, 1997 revealed
that 81% of all patients would prefer early drain removal followed by
outpatient seroma aspiration, if necessary. This preference was also
evident for those patients who actually developed a seroma and required needle aspiration. No patient claimed that it was inconvenient to
return to the hospital. Dalberg et al. (2004) investigated the effect of
seroma formation on patient well-being, general health and functional
status in a subgroup of 82 patients of which fifty percent developed
seroma. No statistically significant differences in QoL were observed.
However, the authors did not mention the method of assessment and
the full questionnaire was not provided.
Several authors mentioned that drains can cause considerable disadvantages for patients such as a longer duration of nursing care, hindrance in daily activities and social life and reduction in mobility but
did not investigate these hypotheses in any further detail (BaasVrancken Peeters et al., 2005; Jeffrey et al., 1995; Taylor et al., 2013;
Zavotsky et al., 1998). Therefore, the goal of our randomised controlled
trial was to compare early drain removal with output-based drain removal in patients who underwent BCS. The primary endpoint was the
drains were removed at the discretion of the patient's physician.
Dalberg et al. (2004) reported a higher incidence of seroma, but the
total volume of fluid collected after percutaneous aspiration did not
differ. These positive effects of drain removal were found in a randomised multi-centre study in five Swedish hospitals. 198 patients with
BC who underwent a modified radical mastectomy with clearance of the
axilla were randomised into two groups: drains were removed one day
after surgery in the first group, while the criterion for the other group
was a flow less than 40 ml/day. Gupta et al. (2001) were the first to
mention a higher number of aspirations and total volume of aspirations
in the early-removal group. They included 121 patients who were
randomised to have drains removed either five or eight days postoperatively (64 and 57 patients, respectively). Only one study has been
published to date that was closed early. This study was halted after
including 24 patients due to higher rates of seroma aspiration, drain
reinsertion and physician visit in the early-removal group (Barton et al.,
2006).
Infection rate is another important endpoint in many studies since
both drains and aspirations can be the cause of infections. Skin bacteria
can enter for example retrograde via the drain and cause infection
(Andeweg et al., 2011). No higher incidence of infection due to early
drain removal has been reported to date. Additional positive effects of
early drain removal are reduction in the length of hospital stay (Dalberg
et al., 2004; Okada et al., 2015; Taylor et al., 2013) and a lower need of
homecare (Andeweg et al., 2011). Ackroyd and Reed, 1997 studied
restrictions in shoulder movement and presence of lymphedema, but
did not find a significant difference between the standard and early
drain removal group.
A systematic review of randomised controlled trials (RCTs) on
timing of drain removal was conducted by Kelley et al. (2012). Data of
RCTs on axillary dissection for different reasons, were collected from
electronic databases such as Medline and Embase. Seven RCTs were
retained, including the study by Ackroyd and Reed (1997), Gupta et al.
(2001), Dalberg et al. (2004). and Baas-Vrancken Peeters et al. (2005).
113
European Journal of Oncology Nursing 36 (2018) 112–118
H. Vos et al.
Table 2
Sample and disease characteristics.
Variable
BMI
Mean ( ± Std)
Median (Range)
Age
Mean ( ± Std)
Median (Range)
Output based (N = 48)
Early removal (N = 51)
P-Value
25.3 ( ± 4.38)
24.6 (18.3; 38.2)
26.6 ( ± 5.13)
26.0 (17.9; 41.3)
0.203
58.3 ( ± 13.85)
57.0 (30.0; 86.0)
56.3 ( ± 12.54)
53.0 (36.0; 90.0)
0.423
N
ASA Score
1
2
3
4
Smoking
No
Yes
Diabetes
No
Yes
Neoadjuvant therapy
No
Yes
Previous radiotherapy
No
Yes
Type of surgery
Mastectomy (+SLNB)
Mastectomy (+SLNB) + ALND
ALND
Surgical technique
Argon
Bipolar
Monopolar
Number of positive lymph nodes
0
1
2
>2
pT
Tis
T1
T2
T3
T4
pN
N0
N1
N2
N3
Tumour type
IDC
ILC
ER+
PR+
Her-2+
%
N
%
10
28
10
0
20.8
58.3
20.8
0.0
8
33
9
1
15.7
64.7
17.7
2.0
0.790
36
12
75.0
25.0
46
5
90.2
9.8
0.062
46
2
95.8
4.2
49
2
96.1
3.9
1.000
36
12
75.0
25.0
41
10
80.4
19.6
0.630
43
5
89.6
10.4
50
1
98.0
2.0
0.105
18
20
10
37.5
41.7
20.8
23
18
10
45.1
35.3
19.6
0.755
6
10
32
12.5
20.8
66.7
11
10
30
21.6
19.6
58.8
0.527
18
12
7
11
37.5
25.0
14.6
22.9
18
16
7
9
36.0
32.0
14.0
18.0
0.854
7
17
14
9
0
14.6
35.4
29.2
18.8
0.0
3
17
25
5
1
5.9
33.3
49.0
9.8
2.0
18
22
3
4
37.5
45.8
6.3
8.3
19
24
3
4
37.3
47.1
5.9
7.8
36
6
43
37
4
75.0
12.5
89.6
77.1
8.3
42
6
46
42
13
82.4
11.8
90.2
82.4
25.5
ALND = axillary lymph node dissection, SLNB = sentinel lymph node biopsy, IDC = invasive ductal carcinoma, ILC, invasive lobular carcinoma, ER = estrogen,
PR = progesterone.
randomised controlled trial. Patients older than 18 who were scheduled
for BCS with placement of suction drainage (mastectomy or ALND)
were included independent of TNM classification. The study was also
presented to patients scheduled for breast-conserving treatment with
sentinel lymph node biopsy (SLNB) to allow inclusion postoperatively
in case an ALND was performed after a positive SLNB. Patients were
randomised with sealed envelopes which were prepared prior to the
start of the study. Combined reconstructive surgery was an exclusion
criterion. Patients in the early-removal group were excluded postoperatively if early drain removal at hospital discharge was not feasible
or not deemed safe by the surgeon, e.g., due to infection or impaired
improvement of postoperative QoL in all its dimensions and the secondary endpoint was the clinical- and safety assessment of early drain
removal.
2. Patients and methods
2.1. Study population
Data of 106 patients who underwent BCS at the multidisciplinary
breast centre (MBC) of the University Hospitals Leuven (UZ Leuven,
Belgium) were collected between September 2015 and March 2016 in a
114
European Journal of Oncology Nursing 36 (2018) 112–118
H. Vos et al.
2.3. Patient related- and clinical outcomes
Table 3
Group comparison for postoperative quality of life.
Question
Output based
(N = 48)
Early removal
(N = 49)
N (%)
N (%)
A questionnaire about patient QoL was conducted pre- and postoperatively. The preoperative questionnaire was presented to the patient the day before surgery and the postoperative questionnaire was
conducted three weeks after surgery. As no validated disease-specific
questionnaire on the role of drains on QoL existed. Study-specific
questionnaires were designed based on patient input, literature review
and input by healthcare professionals specialised in breast cancer
treatment. Interrelations were pointed out between QoL and drain-related problems such as drain-induced pain, discomfort due to seroma or
the drain, sleep disturbance and implications of the presence of the
drain on daily activities and social life.
Six clinical variables were selected in advance to determine safety
and feasibility of early drain removal: incidence of wound infection,
complications in wound healing (defined as unclosed wound 3 weeks
postoperatively), time until start of adjuvant treatment (in days after
surgery), duration of seroma formation (period that drain or seroma
aspiration is needed in days after surgery), total fluid volume (sum of
drain volumes and volume of seroma aspirations if needed) and time
until last wound care appointment at the hospital (in days after surgery). Additionally, the effect of age and body mass index (BMI) on
these six clinical variables was assessed for the full study population,
independent of group classification.
P-Value
1. Do you have any pain complaints at the treated breast since surgery?
No
29 (60.4)
34 (69.4)
0.69
Yes, a little
18 (37.5)
14 (28.6)
Yes, severe
1 (2.1)
1 (2.0)
2. Do you have any pain complaints at the treated shoulder since surgery?
No
31 (64.6)
42 (85.7)
0.043
Yes, a little
14 (29.2)
5 (10.2)
Yes, severe
3 (6.3)
2 (4.1)
3. Do you experience limitations in your daily activities or personal care?
No
10 (20.8)
23 (46.9)
0.003
Yes, a little
25 (52.1)
23 (46.9)
Yes, severe
13 (27.1)
3 (6.1)
4. Do you need extra home nursing/care since surgery?
No
2 (4.2)
37 (75.5)
< .001
Yes, a little
39 (81.3)
10 (20.4)
Yes, severe
7 (14.6)
2 (4.1)
5. Do you experience any disturbance in quality of sleep?
No
14 (29.2)
9 (18.4)
0.24
Yes
34 (70.8)
40 (81.6)
6. Do you experience limitations in your mobility (e.g. walking, cycling or driving a car)?
No
13 (27.1)
34 (69.4)
< .001
Yes, a little
21 (43.8)
13 (26.5)
Yes, severe
14 (29.2)
2 (4.1)
7. Do you feel limited in your daily activities because of frequent wound care?
No
26 (54.2)
45 (91.8)
< .001
Yes, a little
14 (29.2)
3 (6.1)
Yes, severe
8 (16.7)
1 (2.0)
8. Do you feel limited in your social life/activities?
No
23 (47.9)
39 (79.6)
< .001
Yes, a little
18 (37.5)
10 (20.4)
Yes, severe
7 (14.6)
0 (0.0)
2.4. Statistical analysis
wound healing. The study protocol was approved by the local medical
ethics committee of UZ Leuven (S58129) and was registered online
(NCT03104153).
All data analysis was conducted by an expert biostatistician with
SAS software (version 9.4 of the SAS System for Windows). Comparison
of the two treatment groups was performed with the Fisher exact test
for categorical variables or the Mann-Whitney U test for continuous
variables. Variables which were inherently linked to group definition,
i.e., period of home care, period with drainage and number of aspirations were not included in the analysis of significance. Prognostic factors for clinical safety variables were analysed using logistic regression
models for binary variables or linear models for continuous variables.
All tests were two-sided with a five percent significance level
(α = 0.05). Summary statistics are presented as means with standard
deviation (Std) or medians with range for continuous variables and as
frequencies and percentages for categorical variables.
2.2. Trial design
3. Results
Two groups were defined prior to the study. In group 1, the outputbased group, drains were removed when the suction drain flow was less
than 30 ml/day for at least 2 days with no further signs of infection,
collection or impaired wound healing. Ultimately, drains were removed
6–7 weeks postoperatively even if the flow was higher than 30 ml/day.
Leakage or severe patient discomfort led to immediate drain removal at
any time during postoperative care. In group 2, the early-removal
group, drains were removed at hospital discharge 4–5 days after surgery independent of the output at that time. All patients in this study
were allowed to leave the hospital 4–5 days after surgery. Patients had
a weekly wound care appointment after hospital discharge. Clinical
variables such as wound appearance, drain and seroma volumes and
shoulder mobility were recorded. Seroma was defined as a collection of
serous fluid that was clinically visual under the skin flaps or in the
axillary space formed after mastectomy and/or ALND. All mastectomies
and ALND were performed according to UZ Leuven procedure. This
includes Argon-enhanced dissection of the breast and bipolar coagulation for ALND. Two drains were applied prepectoral and/or axillary and
wounds were closed intradermal. All patients were followed from the
day before surgery until their last wound care appointment or until the
start of adjuvant treatment, if applicable. No home care was required in
case of hospital discharge without drains, while this was necessary for
patients with suction drains.
3.1. Exclusion and quality of randomisation
P-values in bold fall below the 0.05 significance level. Two postoperative
questionnaires of patients in the early-removal group were not available for
analysis.
Seven out of 106 patients were excluded, 4 in the output-based
group and 3 in the early-removal group (Fig. 1). Reasons for exclusion
were: incorrect timing of drain removal according to protocol (4 patients), postoperative medical problems affecting the study protocol (1
patient), refusal of the patient to remove drains according to protocol (1
patient) and drain removal not feasible at hospital discharge according
to the surgeon due to severe haematoma (1 patient). Data of the remaining 99 patients (97 female and 2 male, one male in each group)
were taken into account for the statistical analysis. Relevant clinical
features were compared to ensure randomisation was well-balanced
(Table 2).
3.2. Patient related outcomes
The primary endpoint of this study was the improvement of postoperative QoL. A statistically significant difference was noted for six out
of eight relevant questions for the postoperative questionnaire, with the
early-removal group consistently scoring better (Table 3). First, patients
of the early-removal group experienced less shoulder pain after surgery.
Second, they experienced fewer limitations in their daily activities and
felt more mobile than patients of the output-based group. Third,
115
European Journal of Oncology Nursing 36 (2018) 112–118
H. Vos et al.
Table 4
Group comparison for clinical variables indicating safety.
Variable
Infection (n/N (%))
No
Yes
Uncomplicated wound healing (n/N (%))
No
Yes
Time until start adjuvant treatment (days)
N
Mean ( ± Std)
Median (Range)
Output based
Early removal
P-value
42/48
6/48
(87.5)
(12.5)
48/51
3/51
(94.1)
(5.9)
0.31
2/48
46/48
(4.2)
(95.8)
0/51
51/51
(0.0)
(100)
0.23
32
42.4 ( ± 12.3)
42.0 (24.0; 85.0)
37
39.0 ( ± 9.7)
37.0 (21.0; 71.0)
0.18
≤ 6 weeks (n/N (%))
> 6 weeks (n/N (%))
18/32
14/32
(56.3)
(43.8)
27/37
10/37
(72.8)
(27.0)
0.21
≤ 8 weeks (n/N (%))
> 8 weeks (n/N (%))
Duration of seroma formation (days)
N
Mean ( ± Std)
Median (Range)
Total fluid volume (ml)
N
Mean ( ± Std)
Median (Range)
Time until last wound care appointment (days)
N
Mean ( ± Std)
Median (Range)
Number of wound care appointments
N
Mean ( ± Std)
Median (Range)
Number of aspirations
N
Mean ( ± Std)
Median (Range)
Period of home care (days)
N
Mean ( ± Std)
Median (Range)
Period with drainage (days)
N
Mean ( ± Std)
Median (Range)
30/32
2/32
(93.8)
(6.3)
34/37
3/37
(91.9)
(8.1)
1.00
47
26.7 ( ± 12.9)
29.0 (4.0; 51.0)
50
22.4 ( ± 14.3)
20.0 (3.0; 72.0)
0.075
47
1943.9 ( ± 1576.4)
1745.0 (60.0; 6945.0)
50
973.7 ( ± 755.6)
752.5 (47.0; 3285.0)
0.004
47
33.4 ( ± 10.9)
34.0 (20.0; 61.0)
50
29.2 ( ± 10.9)
27.0 (8.0; 72.0)
0.042
47
4.9 ( ± 1.94)
5.0 (2.0; 9.0)
50
4.5 ( ± 2.70)
4.0 (1.0; 14.0)
0.15
47
0.7 ( ± 1.26)
0.0 (0.0; 5.0)
50
3.3 ( ± 3.01)
3.0 (0.0; 14.0)
NR
48
19.3 ( ± 13.00)
16.5 (0.0; 46.0)
51
0.7 ( ± 3.16)
0.0 (0.0; 22.0)
NR
48
22.4 ( ± 13.19)
20.0 (4.0; 48.0)
51
4.6 ( ± 1.59)
4.0 (3.0; 12.0)
NR
P-values in bold fall below the 0.05 significance level. One patient of each group was not taken into account for statistical analysis for some variables due to missing
data. NR = not relevant as inherently linked to group definition.
analysis (32 patients in the output-based group and 37 patients in the
early-removal group). First, group-specific assessment was performed
on the length of the period between surgery and the start of radio- or
chemotherapy. Second, the percentage of patients was calculated that
could start therapy within 6 or 8 weeks after surgery. All patients under
consideration (N = 69) could start their adjuvant treatment within 3
months postoperatively and 93% of the patients started their therapy
within eight weeks after surgery. Only five patients started later than
this eight-week period because of impaired wound healing, the need for
additional testing or additional surgery. None of the group differences
were statistically significant: the median period between surgery and
the start of radio- or chemotherapy was 43 versus 37 days in the advantage of early drain removal. The median period of seroma formation
was 9 days shorter in the early-removal group, albeit not significant.
The number of wound care appointments was not statistically different
in the two groups. However, the total period of home care was 19.3
days on average for the output-based group, in comparison with only
0.7 days for the early-removal group. Furthermore, our data show that
neoadjuvant therapy or a history of radiotherapy of the breast region
does not increase the risk of seroma. Additionally, the data in Table 5
show that both elderly (defined as 70 or older) and patients with a high
BMI (defined as 25 or higher) have a higher risk of negative clinical
patients in the early-removal group required less home care after surgery, accompanied by patient perception of fewer restrictions due to
home care. Furthermore, the new policy of early drain removal was
preferred by 94% of the patients in the early-removal group (N = 48).
No significant difference was found for the question about postoperative pain in the breast region (p-value of 0.69) and the question
about quality of sleep (p-value of 0.24).
3.3. Clinical variables
The secondary endpoint of this study was the clinical- and safety
assessment of early drain removal (Table 4). Two significantly positive
results for the early removal could be observed: lower production of
wound fluid measured as the sum of drain and aspiration volumes
(median 1745.0 ml versus 752.5 ml) and shorter wound care follow-up
(median 34 versus 27 days). Nine patients developed an infection, 6 in
the output-based group and 3 in the early-removal group. Impaired
wound healing occurred in only 2 patients, both in the output-based
group and both of them were smokers. Neither the difference for infection nor for impaired wound healing was statistically significant.
Only data of those patients who were scheduled for adjuvant
treatment at UZ Leuven were included in the adjuvant treatment
116
European Journal of Oncology Nursing 36 (2018) 112–118
H. Vos et al.
Table 5
Statistics for the different clinical variables according to age and body mass index (BMI) classification.
Category
Age
BMI
< 70 (N = 78)
≥70 (N = 21)
P-value
< 25 (N = 47)
≥25 (N = 52)
P-value
Infection (n (%))
No
Yes
74 (94.87)
4 (5.13)
16 (79.19)
5 (23.81)
0.019
44 (93.62)
3 (6.38)
46 (88.46)
6 (11.54)
0.49
Category
Age
< 70 (N = 58)
Time until start adjuvant treatment
Mean ( ± Std)
39.2 ( ± 10.86)
Median (Range)
38.0 (21.0; 85.0)
Category
BMI
≥70 (N = 11)
P-value
< 25 (N = 32)
≥25 (N = 37)
P-value
47.7 ( ± 9.23)
49.0 (28.0; 57.0)
0.005
41.8 ( ± 12.74)
40.0 (21.0; 85.0)
39.5 ( ± 9.31)
40.0 (24.0; 58.0)
0.67
Age
< 70 (N = 77)
Duration of seroma formation
Mean ( ± Std)
21.6 ( ± 11.39)
Median (Range)
20.0 (3.0; 43.0)
Total fluid volume
Mean ( ± Std)
1297.2 ( ± 1257.03)
Median (Range)
915.0 (47.0; 6945.0)
Time until last wound care appointment
Mean ( ± Std)
29.3 ( ± 9.18)
Median (Range)
27.0 (8.0; 50.0)
BMI
≥70 (N = 20)
P-value
< 25 (N = 46)
≥25 (N = 51)
P-value
35.4 ( ± 16.66)
37.5 (5.0; 72.0)
< 0.001
21.2 ( ± 14.08)
20.0 (3.0; 51.0)
27.4 ( ± 12.87)
23.0 (5.0; 72.0)
0.033
2008.4 ( ± 1395.03)
1602.5 (532.0; 4669.0)
0.015
1185.0 ( ± 1083.18)
742.0 (47.0; 4255.0)
1677.3 ( ± 1459.02)
1055.0 (257.0; 6945.0)
0.041
38.7 ( ± 14.34)
37.0 (20.0; 72.0)
0.005
29.8 ( ± 10.91)
27.0 (8.0; 61.0)
32.5 ( ± 11.09)
33.0 (20.0; 72.0)
0.254
The variable “uncomplicated wound healing” was not included in the table due to a low number of events. One patient of each group was not taken into account for
statistical analysis for “Duration of seroma formation”, “Total fluid volume” and " Time until last wound care appointment" due to missing data. P-values in bold fall
below the 0.05 significance level.
members and/or professional cleaning service. In addition, significantly
more patients of the output-based group felt limited in their daily activities because of the necessity of frequent wound care; at home or at
the hospital. This difference cannot be explained by the number of inhospital wound care appointments, since there was no significant difference in the number of these appointments for patients in either of the
two groups. Moreover, several patients mentioned that they perceived
the in-hospital appointments as part of the recovery and not as a great
burden. This allows to state that home care is the only relevant factor in
this context. Patients of the output-based group needed on average 19.3
days of home care by a professional, while patients of the early-removal
group only needed 0.7 days of home care. The latter figure is due to an
older patient who needed 22 days of postoperative home care for hygienic reasons, most patients (N = 47) in the early-removal group did
not have any home care at all. Patients of the output-based group,
however, indicated that they felt physically limited to a great extent
because their social agenda was determined by the home care appointments.
Third, a significant difference was noticed when looking into restrictions in patient mobility with patients in the early-removal group
feeling more mobile. An important aspect in this context is driving,
patients with drains are not allowed to drive a car for insurance reasons.
Not being able to drive is especially inconvenient for singles and patients with children because they become dependent on other people.
Although patients without drains are allowed to drive a car, not all of
them felt comfortable to drive, for example due to limitations in
shoulder mobility.
Finally, patients of the early-removal group felt significantly more
limited in their social life. The presence of a drain kept people at home
because of the visibility of the drains, the limitation in movement and the
daily confrontation with their disease. These results are in contradiction
with a study by Dalberg et al. who found no difference with respect to
QoL. This difference can be attributed to the different approach to assess
QoL, with Dalberg et al. (2004). The positive results with respect to QoL
translate to a positive evaluation of the early-drain removal overall.
outcome. These parameters should not be considered as confounding
factors in this RCT since randomisation was well-balanced as shown in
Table 2.
4. Discussion
Early drain removal was associated with a significant improvement
in QoL since patients experienced considerably less limitations in daily
activities and social life. Additionally, total duration of home care
nursing was lower in the early-removal group, most patients did not
need any home care. Moreover, it was clinically safe to remove drains
early and the total volumes of fluid drained and aspirated were significantly lower in the early-removal group.
4.1. Quality of life
Patients in the early-removal group scored significantly better on
the postoperative questions on QoL.
First, a significant difference was noticed regarding postoperative
shoulder pain, with patients in the early-removal group suffering less
from postoperative pain. This difference could not be explained by the
type of surgery (ALND or SLNB) and might, therefore, be related to the
drain itself.
Second, a significant difference was noticed regarding restrictions in
daily activities and personal care, with patients in the output-based
group experiencing more limitations. Common quoted restrictions in
daily activities are problems with getting dressed, practicing sports and
taking care of household tasks. An important limitation in personal care
is the prohibition of taking a shower or bath with a drain. Similar to the
case with postoperative shoulder pain, these findings cannot be attributed to the type of surgery and are, therefore, related to the drain
itself. Worth mentioning is that patients interpreted question six about
need for extra home care postoperatively broader than only professional home care by a nurse, which is always needed when a drain was
present. Some patients needed extra help from their partner, family
117
European Journal of Oncology Nursing 36 (2018) 112–118
H. Vos et al.
seroma formation and total fluid volume. This finding is in accordance
with Dalberg et al. (2004) and Van Bemmel et al. (2011), both stating
that a high BMI is associated with increased risk of seroma formation.
Care policies for breast cancer patients can be adapted to incorporate
these risk factors both for informing patients and to determine the
frequency of follow-up.
In conclusion, our data show that it is clinically safe and preferable
to remove drains early after breast cancer surgery. Better results were
obtained for patients in the early-removal group, even when statistical
significance was not obtained for each variable. The results of this study
have led to an adaptation of drain-removal policy at our centre, from
output based to early removal at hospital discharge. The authors are
convinced that the findings of this study can be useful for other breast
cancer surgery centres as well, maintaining a case-by-case assessment of
drain policy. Moreover, the potential improvement in QoL for patients
with breast cancer should be rated at its absolute inestimable value.
Three weeks after surgery, patients of the early-removal group were
asked whether they would still prefer early drain removal at discharge
in case they would need a similar intervention in the future. The new
policy of early drain removal was preferred by 94% of these patients.
This number is in close accordance with a study performed by Ackroyd
and Reed, 1997, who found that 81% of the patients preferred early
drain removal even when seroma aspiration after drain removal was
required frequently. The main reasons to prefer early drain removal in
our study were an increased mobility and the lack of social stigma related to the drains.
A limitation of our study is the use of non-validated questionnaires.
As already mentioned, suitable validated alternatives were not available. Study-specific questionnaires were designed to point out interrelations between QoL and drain-related problems. In spite of this
limitation, the significant differences between both groups cannot be
undervalued, indicating a clear effect of the drain-related impact on
QoL for patients with BC after surgery.
Author declaration
4.2. Clinical variables
None declared.
Patients in the output-based group had a significantly higher total
fluid volume. This finding might confirm the theory that drains perpetuate drainage as cited by Andeweg et al. (2011). An additional explanation can be found in the self-absorbing capacity of the body for
wound fluid in the absence of drains. Moreover, the median period of
seroma formation was shorter in the early-removal group, albeit not
statistically significant.
Our results are in close concordance with Dalberg et al. (2004) who
reported a higher incidence of seroma, but the total volume of fluid
collected after percutaneous aspiration did not differ. Only Gupta et al.
(2001) mention a higher number of aspirations and total volume of
aspirations in the early-removal group. The total number of aspirations
(75 in the five day group versus 42 in the eight day group), total cumulative lymphocele volume (13 759 ml versus 6275 ml) and percentage of lymphoceles (48% versus 28%) were statistically higher in the
five day group. The differences in comparison to our study can be partly
explained by the different study design and interpretation of seroma
formation.
The number of patients with a postoperative infection was slightly
higher in the output-based group, albeit not significant. This corresponds with the results of Ackroyd and Reed, 1997, who also observed a
higher percentage of infections in the group with output-based removal
(eight versus five percent). All infections in the output-based group
occurred when the drains were still present, indicating that the presence of drains is indeed associated with an increased risk of infection
as shown by Dalberg et al. (2004) and Baas-Vrancken Peeters et al.
(2005). In theory, seroma aspiration could have been the cause of infection in two patients of the early-removal group in our study. However, this was not the case for the third patient since infection was already present before the first aspiration was performed.
Looking to the data of the full study population, independent of
group classification, our data show that five percent of all patients
younger than 70 developed an infection, while this was 24% in the
group with patients older than 70. These results are in accordance with
a systematic review by Korol et al. (2013), which showed that ageing is
associated with an increased risk of infection. However, there is no
consensus on this association, as stated by Kaye et al. (2005). These
authors determined a more complex relationship between age and infection, with risk of infection increasing linearly until the age of 65 but,
surprisingly, a decrease in infection risk after that age. Similarly, approximately 12% of the patients in our study with a BMI greater than or
equal to 25 developed an infection while this was only 6 percent for
patients with a BMI lower than 25. These results are not surprising as it
is well established in literature that obesity is a major risk factor in the
development of infection (Anaya and Dellinger, 2006; Tjeertes et al.,
2015). Moreover, BMI and age had negative effects on duration of
References
Ackroyd, R., Reed, M.W.R., 1997. A prospective randomized trial of the management of
suction drains following breast cancer surgery with axillary clearance. Breast 6,
271–274. https://doi.org/10.1016/S0960-9776(97)90003-5.
Anaya, D. a, Dellinger, E.P., 2006. The obese surgical patient: a susceptible host for infection. Surg. Infect. 7, 473–480. https://doi.org/10.1089/sur.2006.7.473.
Andeweg, C.S., Schriek, M.J., Heisterkamp, J., Roukema, J.A., 2011. Seroma formation in
two cohorts after axillary lymph node dissection in breast cancer surgery: does timing
of drain removal matter? Breast J. 17, 359–364. https://doi.org/10.1111/j.15244741.2011.01099.x.
Baas-Vrancken Peeters, M.J., Kluit, A.B., Merkus, J.W., Breslau, P.J., 2005. Short versus
long-term postoperative drainage of the axilla after axillary lymph node dissection. A
prospective randomized study. Breast Canc. Res. Treat. 93, 271–275. https://doi.org/
10.1007/s10549-005-5348-7.
Barton, A., Blitz, M., Callahan, D., Yakimets, W., Adams, D., Dabbs, K., 2006. Early removal of postmastectomy drains is not beneficial: results from a halted randomized
controlled trial. Am. J. Surg. 191, 652–656. https://doi.org/10.1016/j.amjsurg.2006.
01.037.
Dalberg, K., Johansson, H., Signomklao, T., Rutqvist, L.E., Bergkvist, L., Frisell, J.,
Liljegren, G., Ambre, T., Sandelin, K., 2004. A randomised study of axillary drainage
and pectoral fascia preservation after mastectomy for breast cancer. Eur. J. Surg.
Oncol. 30, 602–609. https://doi.org/10.1016/j.ejso.2004.03.020.
Gupta, R., Pate, K., Varshney, S., Goddard, J., Royle, G.T., 2001. A comparison of 5-day
and 8-day drainage following mastectomy and axillary clearance. Eur. J. Surg. Oncol.
27, 26–30. https://doi.org/10.1053/ejso.2000.1054.
Jeffrey, S.S., Goodson 3rd, W.H., Ikeda, D.M., Birdwell, R.L., B, M., 1995. Axillary lymphadenectomy for breast cancer without axillary drainage. Arch. Surg. 130, 909–913.
https://doi.org/10.1001/archsurg.1995.01430080111018.
Kaye, K.S., Schmit, K., Pieper, C., Sloane, R., Caughlan, K.F., Sexton, D.J., Schmader, K.E.,
2005. The effect of increasing age on the risk of surgical site infection. J. Infect. Dis.
191, 1056–1062. https://doi.org/10.1086/428626.
Kelley, T.A., Thomson, D.R., Furniss, D., 2012. When should axillary drains be removed
post axillary dissection? A systematic review of randomised control trials. Surg Oncol
21, 247–251. https://doi.org/10.1016/j.suronc.2012.05.002.
Korol, E., Johnston, K., Waser, N., Sifakis, F., Jafri, H.S., Lo, M., Kyaw, M.H., 2013. A
systematic review of risk factors associated with surgical site infections among surgical patients. PLoS One. https://doi.org/10.1371/journal.pone.0083743.
Murphey, B.R., 1947. The use of atmospheric pressure in obliterating axillary dead space
following radical mastectomy. South Surg 13, 372–375.
Okada, N., Narita, Y., Takada, M., Kato, H., Ambo, Y., Nakamura, F., Kishida, A.,
Kashimura, N., 2015. Early removal of drains and the incidence of seroma after breast
surgery. Breast Canc. 22, 79–83. https://doi.org/10.1007/s12282-013-0457-3.
Taylor, J.C., Rai, S., Hoar, F., Brown, H., Vishwanath, L., 2013. Breast cancer surgery
without suction drainage: the impact of adopting a “no drains” policy on symptomatic seroma formation rates. Eur. J. Surg. Oncol. 39, 334–338. https://doi.org/10.
1016/j.ejso.2012.12.022.
Tjeertes, E.E.K.M., Hoeks, S.S.E., Beks, S.S.B.J.C., Valentijn, T.T.M., Hoofwijk, A.A.G.M.,
Stolker, R.J.R.J., 2015. Obesity - a risk factor for postoperative complications in
general surgery? BMC Anesthesiol. 15, 112. https://doi.org/10.1186/s12871-0150096-7.
Van Bemmel, A.J.M., van de Velde, C.J.H., Schmitz, R.F., Liefers, G.J., 2011. Prevention
of seroma formation after axillary dissection in breast cancer: a systematic review.
Eur. J. Surg. Oncol. 37, 829–835. https://doi.org/10.1016/j.ejso.2011.04.012.
Zavotsky, J., Jones, R.C., Brennan, M.B., Giuliano, A.E., 1998. Evaluation of axillary
lymphadenectomy without axillary drainage for patients undergoing breast-conserving therapy. Ann. Surg Oncol. 5, 227–231. https://doi.org/10.1007/BF02303777.
118