European Journal of Cardio-Thoracic Surgery 0 (2017) 1–7 ORIGINAL ARTICLE

doi:10.1093/ejcts/ezx248

Cite this article as: Bjerregaard LS, Jensen PF, Bigler DR, Petersen RH, Møller-Sørensen H, Gefke K et al. High-dose methylprednisolone in video-assisted thoraco-
scopic surgery lobectomy: a randomized controlled trial. Eur J Cardiothorac Surg 2017; doi:10.1093/ejcts/ezx248.

High-dose methylprednisolone in video-assisted thoracoscopic

THORACIC
surgery lobectomy: a randomized controlled trial†
Lars S. Bjerregaarda,b,*, Per F. Jensenb, Dennis R. Biglerb, René Horsleben Petersenc, Hasse Møller-Sørensenb,
Kaj Gefkeb, Henrik J. Hansenc and Henrik Kehleta
a
Section for Surgical Pathophysiology, Rigshospitalet, Copenhagen University, Copenhagen East, Denmark
b
Department of Cardiothoracic Anaesthesia, Rigshospitalet, Copenhagen University, Copenhagen East, Denmark
c
Department of Cardiothoracic Surgery, Rigshospitalet, Copenhagen University, Copenhagen East, Denmark

* Corresponding author. Section for Surgical Pathophysiology, Rigshospitalet, Copenhagen University, Blegdamsvej 9, Section 7621, 2100 Copenhagen, Denmark.
Tel: +45-3545-8946; e-mail: lars.stryhn.bjerregaard@regionh.dk (L.S. Bjerregaard).

Received 7 February 2017; received in revised form 15 May 2017; accepted 18 June 2017

Abstract
OBJECTIVES: The optimal postoperative analgesic strategy after video-assisted thoracoscopic surgery lobectomy remains undetermined.
We hypothesized that high-dose preoperative methylprednisolone (MP) would improve analgesia compared to placebo.
METHODS: A total of 120 adult patients were randomized equally to 125 mg MP or placebo before the start of their elective video-
assisted thoracoscopic surgery lobectomy. Group allocation was blinded to patients, investigators and caregivers, and all patients received
standardized multimodal, opioid-sparing analgesia. Our primary outcome was area under the curve on a numeric rating scale from 0 to
10, for pain scores on the day of surgery and on postoperative days 1 and 2. Clinical follow-up was 2–3 weeks, and telephone follow-up
was 12 weeks after surgery.
RESULTS: Ninety-six patients were included in the primary analysis. Methylprednisolone significantly decreased median pain scores on
the day of surgery: at rest (numeric rating scale 1.6 vs 2.0, P = 0.019) and after mobilization to a sitting position (numeric rating scale 1.7 vs
2.5, P = 0.004) but not during arm abduction and coughing (P = 0.052 and P = 0.083, respectively). Nausea and fatigue were reduced on the
day of surgery (P = 0.04 and 0.03), whereas no outcome was improved on postoperative Days 1 and 2. Methylprednisolone did not in-
crease the risk of complications but increased blood glucose levels on the day of surgery (P < 0.0001).
CONCLUSIONS: High-dose preoperative MP significantly reduced pain at rest and after mobilization to a sitting position on the day of
surgery, without later analgesic effects. Nausea and fatigue were improved without side effects, except transient higher postoperative
blood glucose levels.
CLINICAL TRIAL REGISTRATION: Registered at clinicaltrialsregister.eu [7 November 2012, EudraCT 2012-004451-37; https://www.clini
caltrialsregister.eu/ctr-search/trial/2012-004451-37/DK].
Keywords: Video assisted thaoracic surgery • Pain • Postoperative • Steroids • Methylprednisolone

INTRODUCTION opioid consumption and to enhance recovery after various surgical

The optimal strategy for postoperative analgesia after video-assisted
thoracoscopic surgery (VATS) lobectomy still remains undeter-
mined [1]. In a previous observational study from our institution,
acceptable postoperative analgesia after VATS lobectomy was
achieved using a multimodal analgesic regimen of oral paraceta-
mol, non-steroid anti-inflammatory drugs and gabapentin, supple-
mented by an intraoperative paravertebral nerve block and an
intercostal catheter (ICC) [2]. Preoperative administration of gluco-
corticoids has been shown to reduce acute postoperative pain and
†Presented at the 25th Meeting of the European Society of Thoracic Surgeons,
Innsbruck Austria, 28–31 May 2017.
procedures [3–8] but has not been assessed in VATS lobectomy.
Consequently, we did a double-blinded randomized controlled trial
to investigate the analgesic effect of 125 mg methylprednisolone
(MP) administered preoperatively in patients undergoing VATS lob-
ectomy. Secondarily, we assessed the effects on postoperative nau-
sea, fatigue and sleep quality as well as potential side effects.
MATERIALS AND METHODS
The study was approved by the Regional Ethics Committees of the
Capital Region, Denmark (reference: H-3-2012-148), the Danish
Data Protection Agency (Journal: 2007-58-0015; 30-0845) and the

C The Author 2017. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
V
2 L.S. Bjerregaard et al. / European Journal of Cardio-Thoracic Surgery

Danish Health and Medicine Authority (Journal: 2012103813) under
the control of the Good Clinical Practice unit of Copenhagen
[(www.gcp-enhed.dk/kbh) 25 June 2017, date last accessed] and
was registered with EudraCT (reference: 2012-004451-37). Written
informed consent was obtained from all participants.
We included consenting patients >_18 years who underwent
elective VATS lobectomy at the Department of Cardiothoracic
Surgery, Rigshospitalet, Copenhagen University Hospital, Denmark.
between 21 March 2013 and 23 September 2015. Exclusion crite-
ria were inability to cooperate, inability to speak or understand
Danish, no planned follow-up at Rigshospitalet, allergies to any
used medicines, symptomatic gastric ulceration, New York Heart
Association functional classification >_3, preoperative increased
plasma creatinine, pharmacologically treated diabetes mellitus,
abuse of medicine or alcohol (>_5 units/day), use of psychophar-
macological agents, systemic steroids (except inhalations) and/or
opioid agonists, neurological deficits affecting pain perception,
previous ipsilateral pulmonary resection, pregnancy and/or breast
feeding. Included patients were withdrawn if they were reoperated
on within 72 h, had massive intrathoracic adherences, deviated
from the standard postoperative analgesia, converted to thoracot-
omy and/or breached protocol rules. Patients who received the
study medicine and had completed 48 h of postoperative registra-
tions (with at least 3 timely independent scorings on the day of
surgery) were included in the primary ‘per-protocol’ analysis, even
if they were not available for follow-up. Screening for eligibility
and for participant information was done by an investigator 1–3
days before surgery, allowing time before obtaining written consent.
Before the start of the study, 100 sealed opaque envelopes were
prepared by persons with no relation to the study, based on a
computer-generated randomization sequence (1:1 ratio, block size
20 and no stratification). An additional block of 20 envelopes was
prepared later by the same procedure. Randomization envelopes
and the study medicines were located away from the operating
theatre. When the person was accepted in the study, the investiga-
tor phoned an intensive care unit nurse who opened the next enve-
lope and prepared a 2-ml syringe with 125 mg MP or a 2-ml 0.9%
saline (placebo group), depending on the enclosed information.
The syringe was blinded by a plaster bandage and brought to the
investigator who administered the medicine to the patient after
induction of anaesthesia, but before the start of surgery. Patients,
caregivers and investigators were blinded to group assignment.
Anaesthesia was administered intravenously using propofol, remi-
fentanil and 10 mg of morphine administered about 30 min before
the end of surgery. All patients had a left-sided, double-lumen endo-
tracheal tube, 2 peripheral intravenous catheters and an arterial line.
Pressure-controlled ventilation was used during one-lung ventilation.
Patients were extubated in the operating room and observed in a
specialized postanaesthesia care unit (PACU) until the next morning.
Standard analgesia included paracetamol, ibuprofen and gabapen-
tin, supplemented by a paravertebral nerve block and an ICC placed
intraoperatively by the surgeon [2], with morphine or a bolus of
bupivacaine in the ICC for breakthrough pain. Oral analgesics were
started before surgery and continued for at least 48 h, unless there
were side effects (Table 1). All VATS lobectomies were performed by
experienced thoracic surgeons using a standardized 3-port tech-
nique, as previously described [9, 10].
Preoperative and intraoperative information included demo-
graphic data, American Society of Anesthesiologists physical status
classification, preoperative lung capacity (forced expiratory volume
in 1 s), self-reported preoperative pain (yes/no), the reason for per-
forming VATS lobectomy, duration of surgery and type of resec-
tion. Postoperative pain, nausea and sedation were scored by
trained nurses every 3rd hour from arrival in the PACU until mid-
night on the day of surgery. On postoperative days (POD) 1 and 2,
patients scored pain and nausea twice daily (08:00 and 20:00 h)
according to instructions given by an investigator on the first post-
operative morning. In addition, at 08:00, PODs 1–3, the patients
scored their average fatigue during the last 24 h (on POD 1; since
the end of surgery) and the quality of their sleep the previous
night. Pain was scored on a numeric rating scale (NRS) ranging
from 0 (no pain) to 10 (worst imaginable pain) during 4 prede-
fined activities: at rest, after mobilization from supine to sitting
position (without help), with both arms abducted at least 90 and
when coughing. If a patient reported inability to sit, to abduct the
arm or to cough because of pain, this was interpreted as the ‘worst
imaginable pain’, whereas all other reasons were considered miss-
ing data. For sedation, we used an NRS of 0–4 (0: awake and alert,
1: reacts when spoken to, 2: reacts when touched, 3: reacts to
strong physical stimulation and 4: unresponsive); for nausea, an

Table 1: Standard anaesthesia and analgesia

Preoperative Intraoperative Postoperative From POD 1

Oral PCM 1 g
Oral ibuprofen 800 mg
Oral gabapentine 600 mg
TIVA (propofol/remifentanil)
Paravertebral block by intrathora-
cic surgical technique: (15–20 ml
of 0.5% bupivacaine distributed
equally between intercostal
spaces from Levels T3 to T8)
ICC at end of surgery: continuous
infusion of 0.25% bupivacaine,
6 ml/h
IV PCM 1 g.a morphine 5–10 mg IV
at the end of surgery (up to
15 mg, if BW >_ 90 kg)
Oral PCM 1g  2
Oral ibuprofen 800 mg
Oral gabapentine 300 mg
ICC: continous infusion of 0.25%
bupivacaine, 6 ml/h
For breakthrough pain: 5 ml, 0.25%
bupivacaine in ICC.b Morphine
5–10 mg IV or 10–20 mg orally
Oral PCM 1 g  4
Oral ibuprofen 800 mg  2
Oral gabapentine: 300 mg at 08:00
and 600 mg at 22:00
ICC: continous infusion of 0.25%
bupivacaine, 6 ml/h. Until
removal of chest tube
For breakthrough pain: 5 ml, 0.25%
bupivacaine in ICC.b Morphine
5–10 mg IV or 10–20 mg orally

POD: postoperative day; PCM: paracetamol; TIVA: total intravenous anaesthesia; ICC: intercostal catheter; IV: intravenous; BW: body weight; PACU: postanaes-
thesia care unit.
a
At 12:00, if still in the operating theatre, otherwise administered in the PACU.
b
If the pain was associated with the chest tube.
 L.S. Bjerregaard et al. / European Journal of Cardio-Thoracic Surgery 3

THORACIC
Figure 1: Trial profile. The asterisk indicates each patient may have fulfilled more than one exclusion criteria. ACTH: adrenocorticotropic hormone; ICC: intercostal
catheter; MP: methylprednisolone; NYHA: New York Hearth Association; PO: postoperative; POD: postoperative day; VATS: video-assisted thoracoscopic surgery.

NRS of 0–3 (0: no nausea, 1: light nausea, 2: severe nausea and 3:
vomiting) and for fatigue and sleep quality, an NRS of 1–10 (0: no
fatigue/best imaginable sleep, 10: ‘exhausted’/no or worst imagin-
able sleep). Other postoperative data included use of analgesics
and antiemetics, blood glucose levels on the day of surgery, POD
of removal of the chest tube and ICC and POD of discharge.
Follow-up on complications was obtained in the postoperative
clinic after 2 to 3 weeks and by telephone after 12 weeks.
Our primary outcome was to track pain in the first 48 h post-
operatively. Secondary outcomes included sedation and blood
glucose levels on the day of surgery, use of supplemental anal-
gesics, postoperative nausea, fatigue and sleep quality and the in-
cidence of wound healing problems and/or postoperative
infections within 12 weeks after surgery.
Statistical analysis
In a previous observational study of a similar population, the
mean area under the curve for pain on an NRS 0–10 was 3.2 (SD
1.6) within the first 48 h after VATS lobectomy [2]. We aimed to
detect a pain reduction of 30% (about 1.0 point on the NRS
0–10) with 80% power and a 0.05 significance level (2-sided),
yielding a need for a minimum of 84 patients (42 in each group).
Expecting a dropout rate of 5–10%, we initially included 100 pa-
tients, but as dropouts and exclusions reached about 20% (Fig. 1),
we included an additional 20 patients during the study period.
All analyses were performed ‘per protocol’ by the primary in-
vestigator (L.S.B.) after inclusion of the 120 patients and before
unblinding of data, using the SAS statistical software version 9.4
(SAS Institute, Inc., Cary, NC, USA). Normally distributed continu-
ous data, assessed by the Kolmogorov–Smirnov goodness-of-fit
test, were reported as means with SD and compared between
groups using the unpaired t-test, whereas non-normally distrib-
uted data were reported as medians with interquartile range and
compared between groups by the Mann–Whitney U-test. The
NRS data were considered continuous data and for measure-
ments repeated over time, area under the curves were calculated.
Categorical data are given as counts with group percentages and
compared between groups by the Fisher exact test. Because of
4 L.S. Bjerregaard et al. / European Journal of Cardio-Thoracic Surgery
the different methods of data collection and the different time
intervals between scorings, data from the day of surgery and
from POD 1 and onward were analysed separately. Pain scores
were analysed primarily for those patients with complete data,
but because 28 patients (about 30%) had one or more missing
scores during the day of surgery, we performed a supplemental
multiple imputation model analysis of this period, using age, gen-
der, duration of surgery and the previous pain score as explana-
tory variables.
RESULTS
Of the 422 patients assessed for eligibility, 120 (28%) patients
were included. One (1%) patient withdrew informed consent and
23 (19%) patients were withdrawn by the investigators, leaving 96
(80%) patients for the final analysis; 49 were assigned to the MP
group and 47 to the placebo group (Fig. 1). The median age of
the 302 non-included patients was 69 (interquartile range 62–75)
years; 54% were females. Thirteen of the included patients
(4 from the MP group and 9 from the placebo group) had miss-
ing data on the day of surgery because they arrived in the PACU
after 15:00 h and were therefore not scored after 9 h.
Furthermore, 9 patients were unable to carry out one or more
activities (mobilization, arm spreading and/or coughing) at the
time of arrival in the PACU. Data missing from PODs 1 and 2
were generally fewer and showed no time-specific distributions.
Baseline characteristics and perioperative data are presented
in Tables 2 and 3, and with the exception of gender, these
showed an acceptable balance between the groups. Median pain
scores were lower on the day of surgery in the MP group for all
activities but did not reach statistical significance for arm abduc-
tion and coughing (Table 4). In the multiple imputation analysis,
pain scores were insignificantly reduced during all activities
(P = 0.08, 0.07, 0.18 and 0.22 for rest, arm abduction, mobilization
Table 2: Baseline characteristics for all included and consenting patients
MP (n = 59)
Age (years) 68 (61–75)
Body mass index (kg/m2) 21.2 (4.0); n = 53
Gender
Male 21 (36)
Female 38 (64)
Alcohol >14 units/week 7/53 (13)
Smoking
Current 12/53 (23)
Former 31/53 (58)
Never 10/53 (19)
Patients reporting non-thoracic pain preoperatively 9 (15)
ASA physical status classification
I 3/56 (5)
II 30/56 (54)
III 23/56 (41)
Preoperative FEV1 (% of expected) 86.6 (19.1)
Reason for lobectomy
Confirmed malignancy 46/59 (78)
Suspected malignancy 13/59 (22)
Non-malignant condition 0 3/59 (5)
Data are median (IQR), mean (SD), n (%) or n/N (%).
ASA: American Society of Anesthesiologists; FEV1: forced expiratory volume in 1 s; MP: methylprednisolone; IQR: interquartile range.
and coughing, respectively). Differences in pain scores from
PODs 1 and 2 were generally smaller than on the day of surgery,
without differences in the use of supplemental analgesics at any
time during the postoperative period (Tables 4 and 5). By analy-
sing the pain data for male and females separately, we found no
systematical difference between genders, despite a weak trend
towards higher pain scores in females than males from the MP
group that was not present in the placebo group (Table 4).
Considering all non-imputed pain scores, 24 of 49 (49%) patients
in the MP group and 28 of 47 (60%) patients in the placebo
group reported NRS >_5 at one or more occasions on the day of
surgery (P = 0.31). At PODs 1 and 2, these proportions were 22 of
49 (45%) patients and 29 of 47 (62%) patients for the MP group
and the placebo group, respectively (P = 0.11).
Patients in the MP group had less nausea and fatigue but sig-
nificantly higher blood glucose levels on the day of surgery
(Table 5). No other secondary outcome showed a statistically
significant difference between groups (Table 5).
The incidence of wound healing problems 2–3 weeks postoper-
atively was 5 of 49 (10%) patients in the MP group and 8 of 47
(17%) patients in the placebo group (P = 0.15). In 3 cases, wound
infection was suspected but not proved (2 in MP group and 1 in
placebo group). Twelve weeks after surgery, 2 patients from each
group found their wounds insufficiently healed and 1 patient from
the MP group had received antibiotics for a suspected wound in-
fection. Non-surgical site infections were noted in 19 patients dur-
ing the 12-week follow-up period, with no significant differences
between groups in the type or incidence of these infections.
During the study period, we registered 34 adverse events, of
which 17 were classified as ‘maybe related to study medicine’.
These were 2 cases of new onset postoperative atrial fibrillation
(1 from each group), 1 case of an increased frequency of a pre-
existing atrial fibrillation (MP group), 1 case of transient postop-
erative hypotension (placebo group), 3 cases of stomach pain
(1 from the MP group and 2 from the placebo group), 1 case of
Placebo (n = 60) P-value
66.5 (60–73) 0.17
21.8 (3.9); n = 50 0.50
0.07
32 (53)
28 (47)
10/50 (20) 0.43
0.77
13/50 (26)
26/50 (52)
11/50 (22)
4 (7) 0.15
0.04
8/56 (14)
18/56 (32)
30/56 (54)
85.2 (19.6) 0.68
0.12
38/59 (64)
18/59 (31)
 L.S. Bjerregaard et al. / European Journal of Cardio-Thoracic Surgery 5

transient vision disturbance (placebo group), 2 cases of confusion
(both placebo group), 1 case of increased serum creatinine level
(MP group) and 6 cases who needed insulin because of blood
glucose levels >252 mg/dl on the day of surgery (all in the MP
group). MP did not affect postoperative chest tube duration or
length of stay (Table 3).
surgery. A previous study also assessed the analgesic effect of MP
in open pulmonary resection and found a significant pain reduc-
tion within the PODs 1–2. However, the patients in this study
were given substantially higher doses of MP (25 mg/kg), under-
went pulmonary resection by thoracotomy and received postop-
erative epidural analgesia, which precludes comparison with our

THORACIC
findings [11]. The use of an already well-functioning analgesic re-
gime meant that we generally found low median pain scores in
DISCUSSION both groups in the present study, raising the question of clinical
relevance. However, although they were not all statistically sig-
This study illustrated a moderate additive analgesic effect of a nificant, median pain scores were lower in the MP group for all
single, preoperative dose of 125 mg MP, primarily on the day of activities and time periods and about half of the patients re-
ported severe pain (defined as NRS >_5) [12] at one or more occa-
sions within the first 9 h after surgery. At the same time, we
Table 3: Perioperative characteristics found no increased risk of infection or impaired wound healing,
which is consistent with the general conclusions from recent
MP Placebo P-value publications on perioperative glucocorticoids and postoperative
(n = 49) (n = 47) complications [13–17], thereby supporting the use of glucocortic-
Type of pulmonary resection 0.38 oids for postoperative analgesia after VATS lobectomy. In the
Single lobe 45 (92) 40 (85) present study, we attempted to collect quality data on acute
Single lobe + wedge 4 (8) 5 (11) postoperative pain by setting strict criteria for inclusion, which
resection resulted in exclusion of more than 75% of all eligible patients.
Bilobectomy 0 2 (4)
Duration of surgery (min) 104 (96–118) 102 (88–129) 0.93
Nevertheless, we believe that administration of a single high-
Postoperative day of chest 2 (1–4) 2 (1–5) 0.61 dose of MP may be safe and also have an analgesic effect in the
tube removal majority of patients like those we excluded from this study. Of
Postoperative day of 3 (2–5) 2 (2–5) 0.11 special concern may be the significantly increased blood glucose
discharge levels in the MP group, which resulted in insulin administration
to 6 non-diabetic patients. However, with no registered cases
Data are n (%) or median (IQR). of symptomatic hyperglycaemia, one may question the clinical
MP: methylprednisolone; IQR: interquartile range.
relevance of this transient side effect. Two previous large-scale
studies in cardiac surgery, both of which included patients with

Table 4: Area under the curve for pain scores from patients with complete data

MP (n = 49) Placebo (n = 47) P-value

Pain scores in the first 9 h postoperatively

Pain at rest 1.6 (0.4–2.3); n = 44 2.0 (1.2–3.3); n = 35 0.019
Males 1.3 (0.3–2.3); n = 14 2.5 (1.0–3.3); n = 17
Females 1.8 (0.5–2.3); n = 30 2.0 (1.3–5.2); n = 18
Pain after mobilization to sitting position 1.7 (0.7–2.3); n = 37 2.5 (1.3–4.2); n = 34 0.004
Males 1.5 (0.7–2.0); n = 13 2.8 (1.3–3.9); n = 16
Females 1.8 (0.6–2.5); n = 24 2.4 (1.3–5.2); n = 18
Pain when lifting arms to horizontal position 1.7 (0.8–2.5); n = 39 2.3 (1.2–3.7); n = 31 0.052
Males 1.7 (0.8–2.5); n = 13 2.5 (1.2–3.7); n = 15
Females 1.8 (0.8–2.3); n = 26 2.8 (1.2–4.3); n = 16
Pain when coughing 2.5 (0.9–4.2); n = 40 3.3 (1.8–4.7); n = 32 0.083
Males 1.8 (0.8–4.7); n = 13 3.3 (1.7–4.5); n = 15
Females 2.5 (1.3–3.7); n = 27 3.3 (2.0–4.8); n = 17
Pain scores on PODs 1 and 2
Pain at rest 1.1 (0.5–2.3; n = 46 1.3 (0.7–2.2); n = 43 0.325
Males 0.5 (0–2.2); n = 14 1.3 (0.3–2.0); n = 22
Females 1.3 (0.5–2.3); n = 32 1.5 (0.8–2.2); n = 21
Pain after mobilization to sitting position 1.2 (0.5–2.2); n = 47 1.7 (1.0–2.8); n = 41 0.113
Males 1.0 (0–1.8); n = 15 1.6 (1.2–2.0); n = 22
Females 1.3 (0.8–2.9); n = 32 2.0 (0.8–3.0); n = 19
Pain when lifting arms to horizontal position 1.0 (0.5–1.8); n = 47 1.2 (0.5–2.5); n = 43 0.311
Males 0.9 (0.2–1.3); n = 16 1.2 (0.7–2.3); n = 23
Females 1.0 (0.5–1.8); n = 31 0.9 (0.4–2.8); n = 20
Pain when coughing 2.2 (1.7–4.5); n = 46 3.3 (2.3–5.0); n = 43 0.116
Males 2.0 (1.0–3.3); n = 16 3.3 (2.7–5.7); n = 23
Females 2.4 (1.8–5.0); n = 30 3.3 (1.7–4.0); n = 20

All data are median (IQR).

MP: methylprednisolone; POD: postoperative day; IQR: interquartile range.
6 L.S. Bjerregaard et al. / European Journal of Cardio-Thoracic Surgery

Table 5: Secondary outcomes

MP (n = 49) Placebo (n = 47) P-value

Supplemental analgesia
POD 0
Opioid equivalents (mg of IV morphine) 5 (0–12.5) 5 (0–15) 0.69
Number of boluses in the IC catheter 2 (1–4) 3 (1–5) 0.25
POD 1
Opioid equivalents (mg of IV morphine) 0 (0–5) 0 (0–5) 0.95
Number of boluses in the IC catheter 1 (0–3) 1 (0–2) 0.96
POD 2
Opioid equivalents (mg of IV morphine) 0 (0–0) 0 (0–0) 0.95
Number of boluses in the IC catheter 0 (0–1); n = 29 0 (0–0.5); n = 24 0.24
Postoperative sedation (NRS 0–4)
First 9 h postoperatively (AUC) 0 (0–0.17); n = 44 0 (0–0.17); n = 36 0.46
Blood glucose levels (mg/dl)
First 9 h postoperatively (AUC) 168 (155–189); n = 40 118 (109–130); n = 35 < 0.0001
Nausea (NRS 0–3)
First 9 h postoperatively (AUC) 0 (0–0) [0–0.33]; n = 44 0 (0–0) [0–1.0]; n = 36 0.04
From 8:00, POD 1 to 20:00, POD 2 (AUC) 0 (0–0); n = 43 0 (0–0.17); n = 41 0.51
Number of patients needing antiemetic drugs
POD 0 4 (8) 11 (23) 0.05
POD 1 4 (8) 4 (9) 1.00
POD 2 2 (4) 2 (4) 1.00
Postoperative fatigue (NRS 0–10)
From end of surgery to 8:00, POD 1 3 (1–5) 5 (3–7) 0.03
From 8.00, POD 1 to 08:00, POD 2 3 (1–4); n = 47 3 (1–5); n = 45 0.82
From 8.00, POD 2 to 08:00, POD 3 3 (1–4); n = 47 3 (1–5; n = 42 0.45
Postoperative sleep quality (NRS 0–10)
First postoperative night 5 (1–9) 5 (1–8) 0.25
Second postoperative night 2 (1–5); n = 46 2 (1–5); n = 45 0.83
Third postoperative night 3 (1–5); n = 47 3 (1–5); n = 41 0.81

Data are n (%), n/N (%), median (IQR) or median (IQR) [range].
AUC: area under curve; IV: intravenous; IC: intercostal; MP: methylprednisolone; NRS: numeric rating scale (see text for details); POD: postoperative day.

diabetes, also found increased blood glucose levels after adminis-
tration of glucocorticoids but without relation to any clinically
relevant adverse events, that would contraindicate their use in
patients with or without diabetes [15, 16]. Consistent with results
from previous studies [4, 7, 18], we observed significantly less fa-
tigue and nausea in the MP group on the day of surgery, and al-
though low-point estimates of the latter may indicate little and
not severe nausea in both groups, one should also consider the
difference in the antiemetic drugs used (Table 5). We could not
confirm the previously suggested impairment of postoperative
sleep quality after MP administration [4].
In the present study, we used a standard dose of 125 mg MP,
which is less than that reported in other procedures [8, 11, 15, 16]
but which has been shown to have a pain-reducing effect in
arthroplasty [4, 5] and breast surgery [7]. The analgesic effect has
previously been suggested in response to high doses of 1–2 mg/kg
MP or 0.2–0.4 mg/kg dexamethasone [19] or of ‘intermediate’
doses of 0.11–0.2 mg/kg dexamethasone [3, 6, 14]. However, with
only a few well-defined procedure-specific dose-finding studies
available, these findings cannot conclusively be extrapolated to
specific procedures. Our findings suggest that the analgesic effect
of a single preoperative dose of MP may be of limited duration
after VATS lobectomy, thereby raising the question of repeated
MP administration in the postoperative period to obtain a sus-
tained pain reducing effect [20]. Therefore, more dose–response
and duration studies are needed to elucidate the optimal dose of
MP to reach an analgesic effect but at the same time minimizing
potential side effects, such as hyperglycaemia. Although this is not
fully understood, the pain-reducing effect of glucocorticoids is
believed to be mediated through a general attenuation of the sys-
temic inflammatory response [21]. Also, it has been postulated that
systemic glucocorticoids may prolong the duration of peripheral
nerve blocks [22], which in our study would mean a potentially
prolonged effect of the intraoperatively placed paravertebral nerve
block. Nevertheless, additional studies are needed to elucidate the
underlying biochemical mechanisms of glucocorticoids to further
facilitate their optimal use in reducing acute postoperative pain.
Limitations
This study was limited because the observed pain scores in the
placebo group were lower than we expected, thereby introducing
the risk of a type II error. Also, our power calculation was based
on a mean AUC for pain within a 48-h period, whereas data col-
lection and reporting were done separately from the day of the
operation and from PODs 1 and 2, respectively. Furthermore, we
had a considerable amount of missing pain data from the day of
the operation, which was unequally balanced between the 2 arms.
As reported in the results section, many data were missing because
of operations performed late in the day. Therefore, we had no
pain scores 9 h postoperatively. Assuming that one cannot predict
pain based on the time of the operation, these data may be con-
sidered as ‘missing at random’. We cannot rule out that missing
data related to pain during activities may introduce bias if, for
example, the lack of pain can be explained by sedation, which was
 L.S. Bjerregaard et al. / European Journal of Cardio-Thoracic Surgery
why we performed the ’sensitivity analysis’ with multiple imputa-
tions of missing data. Another potential bias was that neither care-
givers nor investigators were blinded to blood glucose levels,
thereby theoretically unmasking group allocation in the few
patients with very high glucose levels. Finally, our strict inclusion
criteria resulted in a low inclusion rate, which may compromise
the generalizability of our findings. The double-blinded, random-
ized design strengthens the study, and the fact that it was con-
ducted in a single centre means that all patients were operated on
by the same surgical technique, were treated under similar perio-
perative settings, and were followed by just a few investigators,
thereby minimizing protocol violations. Also, we assessed pain
both at rest and under well-defined activities, which may be rele-
vant in relation to postoperative recovery and the potential devel-
opment of chronic pain after thoracic surgery [1, 23]. Furthermore,
we included follow-up visits at 2 separate points to assess potential
complications both in the ’early’ post-discharge period (2-3 weeks
postoperatively) and in the subsequent period, during which most
patients had returned to normal daily activities (3 weeks to 3
months postoperatively).
CONCLUSION
In conclusion, we found that high-dose MP administered preop-
eratively significantly reduced pain on the day of surgery, both at
rest and after mobilization to a sitting position, and reduced nau-
sea and fatigue but had no lasting effects on PODs 1 and 2.
Except for transient-increased blood glucose levels on the day of
surgery, we found no indication of an increased risk of complica-
tions in the MP group.
ACKNOWLEDGEMENTS
We wish to thank Kim Wildgaard for his help and guidance in
preparing the study protocol. Also, a great thanks to the person-
nel from the intensive care unit and the ward in the
Department of Cardiothoracic Surgery, Rigshospitalet, for help-
ing with patient inclusion and data collection.
Funding
This work was supported by the ‘Doctor Fritz Karner and wife
Edith Karner’s Foundation’ and the ‘Doctor Edgar Schnohr and
wife Gilberte Schnohr’s Foundation’.
Conflict of interest: L.S.B., P.F.J., D.B., H.M-S., K.G and H.K. de-
clare no conflicts of interest. R.H.P. is a speaker for Medtronic,
Ethicon and Medela. H.J.H. is a speaker for Covidien/Medtronic,
Bard and Medela.
REFERENCES
[1] Steinthorsdottir KJ, Wildgaard L, Hansen HJ, Petersen RH, Wildgaard K.
Regional analgesia for video-assisted thoracic surgery: a systematic re-
view. Eur J Cardiothorac Surg 2014;45:959–66.
[2] Wildgaard K, Petersen RH, Hansen HJ, Moller-Sorensen H, Ringsted TK,
Kehlet H. Multimodal analgesic treatment in video-assisted thoracic
7
surgery lobectomy using an intraoperative intercostal catheter. Eur J
Cardiothorac Surg 2012;41:1072–7.
[3] De Oliveira GSJ, Almeida MD, Benzon HT, McCarthy RJ. Perioperative
single dose systemic dexamethasone for postoperative pain: a meta-
analysis of randomized controlled trials. Anesthesiology 2011;115:
575–88.
[4] Lunn TH, Kristensen BB, Andersen LO, Husted H, Otte KS, Gaarn-Larsen
THORACIC
L et al. Effect of high-dose preoperative methylprednisolone on pain
and recovery after total knee arthroplasty: a randomized, placebo-
controlled trial. Br J Anaesth 2011;106:230–8.
[5] Lunn TH, Andersen LO, Kristensen BB, Husted H, Gaarn-Larsen L,
Bandholm T et al. Effect of high-dose preoperative methylprednisolone
on recovery after total hip arthroplasty: a randomized, double-blind,
placebo-controlled trial. Br J Anaesth 2013;110:66–73.
[6] Nielsen RV, Siegel H, Fomsgaard JS, Andersen JD, Martusevicius R,
Mathiesen O et al. Preoperative dexamethasone reduces acute but not
sustained pain after lumbar disk surgery: a randomized, blinded,
placebo-controlled trial. Pain 2015;156:2538–44.
[7] Romundstad L, Breivik H, Roald H, Skolleborg K, Haugen T, Narum J
et al. Methylprednisolone reduces pain, emesis, and fatigue after breast
augmentation surgery: a single-dose, randomized, parallel-group study
with methylprednisolone 125 mg, parecoxib 40 mg, and placebo. Anesth
Analg 2006;102:418–25.
[8] de la Motte L, Kehlet H, Vogt K, Nielsen CH, Groenvall JB, Nielsen HB
et al. Preoperative methylprednisolone enhances recovery after endo-
vascular aortic repair: a randomized, double-blind, placebo-controlled
clinical trial. Ann Surg 2014;260:540–8.
[9] Hansen HJ, Petersen RH, Christensen M. Video-assisted thoracoscopic
surgery (VATS) lobectomy using a standardized anterior approach. Surg
Endosc 2011;25:1263–9.
[10] Hansen HJ, Petersen RH. Video-assisted thoracoscopic lobectomy using
a standardized three-port anterior approach—the Copenhagen experi-
ence. Ann Cardiothorac Surg 2012;1:70–6.
[11] Bigler D, Jonsson T, Olsen J, Brenoe J, Sander-Jensen K. The effect of pre-
operative methylprednisolone on pulmonary function and pain after
lung operations. J Thorac Cardiovasc Surg 1996;112:142–5.
[12] Gerbershagen HJ, Rothaug J, Kalkman CJ, Meissner W. Determination of
moderate-to-severe postoperative pain on the numeric rating scale: a
cut-off point analysis applying four different methods. Br J Anaesth
2011;107:619–26.
[13] Wang AS, Armstrong EJ, Armstrong AW. Corticosteroids and wound
healing: clinical considerations in the perioperative period. Am J Surg
2013;206:410–7.
[14] Waldron NH, Jones CA, Gan TJ, Allen TK, Habib AS. Impact of periopera-
tive dexamethasone on postoperative analgesia and side-effects: system-
atic review and meta-analysis. Br J Anaesth 2013;110:191–200.
[15] Dieleman JM, Nierich AP, Rosseel PM, van der Maaten JM, Hofland J,
Diephuis JC et al. Intraoperative high-dose dexamethasone for cardiac
surgery: a randomized controlled trial. JAMA 2012;308:1761–7.
[16] Whitlock RP, Devereaux PJ, Teoh KH, Lamy A, Vincent J, Pogue J et al.
Methylprednisolone in patients undergoing cardiopulmonary bypass
(SIRS): a randomised, double-blind, placebo-controlled trial. Lancet
2015;386:1243–53.
[17] Toner AJ, Ganeshanathan V, Chan MT, Ho KM, Corcoran TB. Safety of
perioperative glucocorticoids in elective noncardiac surgery: a system-
atic review and meta-analysis. Anesthesiology 2017;126:234–48.
[18] Kehlet H. Glucocorticoids for peri-operative analgesia: how far are we from
general recommendations? Acta Anaesthesiol Scand 2007;51:1133–5.
[19] Romundstad L, Stubhaug A. Glucocorticoids for acute and persistent
postoperative neuropathic pain: what is the evidence? Anesthesiology
2007;107:371–3.
[20] Romundstad L, Breivik H, Niemi G, Helle A, Stubhaug A.
Methylprednisolone intravenously 1 day after surgery has sustained anal-
gesic and opioid-sparing effects. Acta Anaesthesiol Scand 2004;48:1223–31.
[21] Holte K, Kehlet H. Perioperative single-dose glucocorticoid administra-
tion: pathophysiologic effects and clinical implications. J Am Coll Surg
2002;195:694–712.
[22] Abdallah FW, Johnson J, Chan V, Murgatroyd H, Ghafari M, Ami N et al.
Intravenous dexamethasone and perineural dexamethasone similarly
prolong the duration of analgesia after supraclavicular brachial plexus
block: a randomized, triple-arm, double-blind, placebo-controlled trial.
Reg Anesth Pain Med 2015;40:125–32.
[23] Wildgaard K, Ravn J, Kehlet H. Chronic post-thoracotomy pain: a critical
review of pathogenic mechanisms and strategies for prevention. Eur J
Cardiothorac Surg 2009;36:170–80.