Clinical Neurology and Neurosurgery 122 (2014) 1219

Contents lists available at ScienceDirect

Clinical Neurology and Neurosurgery

journal homepage: www.elsevier.com/locate/clineuro

Inuence of smoking status at time of surgery for herniated lumbar

disk on postoperative pain and health-related quality of life
Martin N. Stienen a, *, Nicolas R. Smoll b , Gerhard Hildebrandt a , Karl Schaller b ,
Oliver P. Gautschi a,b
a
b

Department of Neurosurgery, Kantonsspital St. Gallen, St. Gallen, Switzerland

Department of Neurosurgery and Faculty of Medicine, University Hospital of Geneva, Geneva, Switzerland

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 16 September 2013
Received in revised form 26 February 2014
Accepted 13 April 2014
Available online 22 April 2014

Background: It is well established that smoking has a myriad of negative effects on varies aspects of bodily
health. The aim of this study was to examine the effects of the smoking status at time of surgery on the
postoperative subjective pain course and health related quality of life (HRQoL) until 1 year after surgery
for lumbar disc herniation (LDH).
Method: This prospective cohort study included patients 18 and 90 years of age with a symptomatic
and radiological veried LDH. The current smoking patient collective (smoking 1 or more cigarettes a
day) was compared with the nonsmoking collective (previous smokers without cigarette consumption
for >2 months and never smokers) in respect of subjective pain sensation (measured with the visual
analogue scale (VAS)) and HRQoL using the short-form (SF-12) questionnaire preoperatively, before
discharge, as well as after 4 weeks and 1 year postoperatively. The primary outcome measures were the
1-year SF-12 scores (MCS and PCS) categorized into responders and non-responders.
Results: A total of 102 patients were enrolled in the study. Thirty-eight patients were current smokers
(37.2%), whereas 43 (42.2%) and 21 (20.6%) patients were never-smokers and previous smokers,
respectively. Four weeks and one year after surgery, both smokers and nonsmokers reported increase in
the HRQoL as compared to preoperative values the MCS increased more than the PCS. From a univariate
and multivariate perspective, smoking status at time of surgery did not predict responder status.
Conclusions: The present study results could not conrm the hypothesis that smoking at time of surgery
was associated with worse outcome after surgery for LDH.
2014 Elsevier B.V. All rights reserved.

Keywords:
Smoking
Lumbar disc herniation
Functional outcome
Lumbar spine surgery
Health related quality of life
Multiple imputation

1. Introduction
Smoking is known to be associated not only with many chronic
diseases, but also with substantial burdens on healthcare systems
worldwide [1]. Furthermore, it is well established that smoking has
a negative effect on postoperative outcome for patients undergoing
surgical procedures including a higher likelihood of 30 day
mortality and serious postoperative complications [1,2]. Smokers
are also at greater risk of postoperative wound healing disturbances compared to nonsmokers [3]. This information is of
paramount importance for the patient information and informed
consent prior to elective surgery, as there is evidence that

* Corresponding author at: Department of Neurosurgery Kantonsspital St. Gallen

Rorschacher Strasse 95, 9007 St. Gallen, Switzerland. Tel.: +41 71 494 1111; fax: +41
71 494 2883.
E-mail address: mnstienen@gmail.com (M.N. Stienen).
http://dx.doi.org/10.1016/j.clineuro.2014.04.015
0303-8467/ 2014 Elsevier B.V. All rights reserved.

perioperative smoking cessation may reduce peri- and postoperative complications [4,5]. Therefore, the patient has to be informed
about the negative effect of smoking in the peri- and postoperative
phase and about a possible advantageously effect of smoking
cessation. The subsequent benet of smoking cessation may not
only be apparent in the immediate postoperative recovery but also
in the long-term convalescence [6].
Although there is growing amount of literature across many
surgical specialties (e.g. in the area of breast surgery, orthopedic
and reconstructive surgery, obstetric surgery, general surgery,
cardiac surgery, head and neck surgery, lung cancer surgery)
regarding the effects of smoking on short- and long-term
complications, cancerogenity, respiratory and cardiovascular
morbidity, reduced health-related quality of life (HRQoL), and
premature death on a increasing number of diseases, there is
dearth of literature on the effect of smoking on the postoperative
pain course and HRQoL in lumbar spine surgery. Still, it has been
demonstrated that spinal fusion rates are signicantly worse in

M.N. Stienen et al. / Clinical Neurology and Neurosurgery 122 (2014) 1219

smokers with nonunion occurring in 40% of cases and 5 times more

often than in nonsmokers (8%) [7]. Interestingly, an improved
spinal fusion rate could be achieved by preoperative nicotine and
smoking cessation in an animal model [8]. Clinical data indicates
that even postoperative smoking cessation leads to better fusion
rates in patients after fusion surgery [9].
According to the above-mentioned literature and our clinical
observation we intended to test the hypothesis that smoking
patients experience more peri- and postoperative complications,
suffer more pain and improve less until one year after surgery for
lumbar disc herniation (LDH).
2. Methods and materials
The study was conducted as a prospective cohort study,
including patients with a symptomatic and radiological veried
LDH. Patients eligible for the study were 18 and 90 years of age
and were operated between October 2010 and February 2011 in our
department. Study exclusion criteria were preoperative systemic
sepsis, disseminated cancer disease, open wound infection before
surgery, or known bleeding disorders. Institutional review board
approval was obtained from the local ethics committee. After
written informed consent, all patients received a specially
designed baseline questionnaire concerning demographic and
social case history, education and profession information, diagnosis and bodily restrictions, detailed pain data, cardiovascular risk
factors, and detailed smoking habits. Before hospital discharge, all
patients lled out a second questionnaire containing diagnosis and
bodily restrictions as well as a detailed pain data sheet. A third and
fourth questionnaire concerning social data, current profession,
diagnosis and bodily restrictions, detailed pain data, cardiovascular risk factors, and a detailed smoking evaluation was mailed to all
patients 4 weeks and 1 year postoperatively, respectively. The
subjective pain sensation was estimated via a standardized
questionnaire containing the visual analogue scale (VAS). The
HRQoL was assessed by the German version of the 12-item shortform health survey (SF-12) questionnaire [10]. The perioperative
complication rate was evaluated by analyzing the incidence of
incidental durotomies, wound infections, and need for
re-operation as well as by the estimated blood loss (EBL) during
surgery, the operation time and the length of overall hospitalization time.
2.1. Study groups
Current smokers were dened as smokers who were smoking
one or more cigarettes a day. Previous smokers were dened as
persons who did not smoke for at least 2 months prior to the
neurosurgical intervention. Finally, never smokers were dened as
those who never smoked in the previous year and reported
0 lifetime pack-years (PY). For the analysis, previous smokers and
patients who never smoked were combined and compared with
the smoking patients at the time of the hospitalization. The choice
of grouping non-smokers and previous smokers was based on
previous studies use of smoking status at surgery [9,1122]. In
addition, this grouping was preferred because of the nding in
other studies that smoking status at surgery was better at
predicting outcomes than cumulative smoking status [9,12].
2.2. Statistical methods
Balance in baseline and radiographic variables was evaluated
using Fishers exact tests for categorical variables. The presence of
missing data necessitated the use of multiple imputation to
complete the dataset in order to assess inuence of the missing
values on standard errors. After the results were imputed (see next

13

section) the baseline mental component scores (MCS) and physical

component scores (PCS) were subtracted from the patients
respective 1-year MCS and PCS score. The results were categorized
into the previously validated ve-tier categories for LBP (much
better, better, little better, same, a little worse and worse)
according to Luo et al., [23] and then turned into a binary
responder and non-responder variable. Responders were dened
as patients feeling much better, better and a little better. In
contrast, non-respondence was determined as no change or
worsening compared to preoperatively. Primary outcome therefore was dened SF-12 scores at 1-year postoperatively categorized
into responders and non-responders. From here, multivariable
logistic regression models were built to assess the relationship of
smoking to outcomes in the presence of potential confounders. For
this, a univariate perspective and a multivariate model was used
with full entry methods of model building [24]. Results were
presented in standard regression tables with additional estimates
of the percentage change of standard errors due to the missing
data. The software used for the statistical analysis was Stata v11.2
(College Station, Texas).
2.3. Multiple imputation
Multiple imputation was used to handle the missing data load
of 34% at 1-year follow-up. Essentially, multiple imputation allows
us to maintain the integrity of the cohort by viewing missing values
as a source of variability to be averaged over, rather than just
omitting data, and therefore effectively ruining the cohort.
Multiple imputation requires three steps. The rst step estimates
the values for the missing data using regression methods (a model
which includes random variation) on the basis of all the present
information in the dataset, which allows easy reproducibility. The
dataset is reproduced m times, so that there are m + 1 datasets.
From here, a standard complete-data model is created (in this case
a logistic regression model), and run on each imputation including
the original (m + 1 times). Then, the estimates are combined to
present within imputation variance and between imputation
variance (standard error) estimates. It is this part of multiple
imputation which is perhaps the most valuable to the read. Within
imputation, variance represents the standard variation, and the
between variance estimates the added variance which is due to the
missing values. This assumes that the data are missing at random,
which means amongst other things that the variable with missing
data itself cannot predict the probability of the data being missing.
In other words, 1-year SF-12 scores (variable with missing data in
this dataset) cannot predict the probability of these values being
missing. We imputed 4 variables (Y): namely the SF-12 PCS as well
as the SF-12 MCS at 1-year postoperatively, as well as the changes
in these scores when compared to baseline (PCS/MCS 1-year
postoperatively minus PCS/MCS at baseline). The number of
imputations (m) was set at 35.
3. Results
3.1. Baseline demographics and neurological status
A total of 112 patients were screened between October 2010 and
February 2011; six refused consent. Four patients had to be
excluded because their health status increased under conservative
therapy and surgery was therefore not performed. After considering exclusion criteria, a total of 102 patients were enrolled in the
study, with all 102 patients undergoing surgery for LDH. Thirtyeight patients were current smokers (37.2%), whereas 43 (42.2%)
and 21 (20.6%) patients were never-smokers and previous
smokers, respectively. Baseline demographic data and the preoperative status are depicted in Table 1. Besides a higher prevalence of

14

M.N. Stienen et al. / Clinical Neurology and Neurosurgery 122 (2014) 1219

Table 1
Besides a higher prevalence of sensory differences in smoking patients, baseline demographic properties and the preoperative status were similar in smoking and
nonsmoking patients. Smoking patients tended to be younger and were therefore less likely to be retired however, differences between the groups were insignicant.
ADL = activity of daily living; VAS = visual analogue scale; MCS = mental component score; PCS = physical component score; SD = standard deviation.
Nonsmokers

Smokers

Total

p-value

Age (years, mean  SD)

53.1  14.9

47.9  11.7

51.2  14.0

0.071

Sex
Male
Female

40
24

62%
38%

26
12

68%
32%

66
36

65%
35%

0.669

Tertiary education
No
Yes
Education in years (mean  SD)
Currently jobless
Disability payments

49
15
12.1  3.5
1
6

77%
23%

84%
16%

0.853

11%
13%

81
21
11.7  3.9
5
11

79%
21%

2%
9%

32
6
10.9  4.4
4
5

Retirement status
Retired
Unretired

14
50

22%
78%

3
35

8%
92%

17
85

17%
83%

0.067

Relationship status
Single/alone
Married/partners

17
46

27%
73%

11
27

29%
71%

28
73

28%
72%

0.831

Previous spine surgery

No
Yes

48
16

75%
25%

25
13

66%
34%

73
29

72%
28%

0.367

Number of previous lumbar surgeries

0
1
2

48
13
3

75%
20%
5%

25
7
6

66%
18%
16%

73
20
9

72%
20%
8%

0.326

Neurological status
Motor decits
Sensory decits
Dependence in ADL
Daily pain medication
VAS pain (mean  SD)

37
34
49
39
5.5  2.6

58%
53%
77%
61%

18
29
31
25
6.3  2.0

47%
76%
82%
66%

55
63
80
64
5.8  2.4

54%
62%
78%
63%

0.306
0.020
0.625
0.676
0.104

SF-12 baseline
MCS (mean  SD)
PCS (mean  SD)

43.9  12.4
31.2  10.2

Total

64

39.5  11.4
30.5  7.8
100%

sensory decits in smoking as compared to nonsmoking patients

(76.3% vs. 53.1%, p = 0.022), all other potential confounders were
evenly balanced between the study groups. Generally, both mean
PCS and MCS values obtained before surgery were in line with the
reference levels described for a German-speaking collective with
LBP, demonstrating that our obtained SF-12 results were stable
[10]. All surgery-related factors were equally distributed between
the study groups (Table 2).
3.2. Missing data review and generation of the multiple imputation
dataset
Follow-up was achieved in 102 patients postoperatively, in
94 patients after 4 weeks and in 67 patients 1-year after surgery.
Known reasons for loss of follow-up were death unrelated to the
surgery in two patients and relocation in ve cases. Interestingly,
smokers were three times more likely to have missed their oneyear follow-up (odds ratio (OR) of 3), and the mechanism
responsible for the missing data was considered to be the missing
at random mechanism. Rather than use the case deletion method,
effectively ruining the cohort and potentially biasing our results,
multiple imputation with 35 imputations was employed for four
variables: PCS, MCS, PCS change from baseline, and MCS change
from baseline. We included smoking status, age, sex, education,

38

5%
11%

42.3  12.2
30.9  9.3
100%

102

0.118
0.096
0.552

0.077
0.718
100%

100%

disability payments, and work status in our multiple imputation

models. Fig. 1 is a density histogram for the observed (complete)
values vs. the imputed values of the 35 imputed datasets (m = 35).
The imputed datasets followed a very similar distribution when
compared to the observed data, except with more variation (see
larger spread of the data).
3.3. Report of postoperative pain and HRQoL in smokers and
nonsmokers
There were no signicant uniform differences in the postoperative average or maximal pain and the SF-12 HRQoL-scores
between smokers and nonsmokers before discharge, at 4 weeks
and 1 year. Even though smoking patients showed a tendency
towards higher VAS pain values and reduced HRQoL (with greater
difference in the SF-12 MCS as compared to the SF-12 PCS), this
observation was short of signicance, except for the SF-12 MCS
1-year after surgery.
Even though average and maximal pain levels decreased
immediately postoperatively in patients of both groups (mean
reduction in average VAS of 3.36 points in smokers and of 2.75
points in nonsmokers; mean reduction in maximum VAS of 2.76
points in smokers and of 2.94 points in nonsmokers), at this point a
marked increase of the HRQoL as measured by the SF-12 MCS and

M.N. Stienen et al. / Clinical Neurology and Neurosurgery 122 (2014) 1219
Table 2
Variables concerning the surgical procedures performed in smokers and
nonsmokers. There were no differences concerning intra- or perioperative
complications between smokers and nonsmokers. CEF = cerebrospinal uid;
min = minutes; ml = milliliters; SD = standard deviation.
Nonsmokers Smokers

Total

Spinal levels with surgery

One
Two

58
6

91%
9%

35
3

92%
8%

93
9

91% >0.99
9%

Surgical approach
Interlaminar fenestration
Foraminotomy
Endoscopic surgery

48
12
4

75%
19%
6%

30
6
2

76%
18%
6%

78
18
6

79% 0.810
16% 0.793
5%

Extend of surgery
Sequestrectomy only
Nucleotomy/sequestrectomy
Other

26
37
1

41%
58%
1%

11
25
2

29%
66%
5%

37
62
3

36% 0.289
61% 0.530
4%

Intraoperative complications
CSF leak
Vascular injury
Deep wound infection
Rebleeding
No complications
Surgery time in min
(mean  SD)
Estimated blood loss in ml
(mean  SD)
Postoperative complications
No complications
Reoperation
Discharge
Home
Rehabilitation clinic
Length of postoperative
hospitalization in days
(mean  SD)
Total

p-value

2
3%

0%

0%

0%
62
97%
73.6  33.5

2
5%

0%

0%

0%
36 95%
73.9  40.6

83.2  14.6

72.9  24.8 78.9  13.3 0.707

63
1

38

98%
2%

100%
0%

4
4%

0%

0%

0%
98
96%
73.7  36.3

101
1

0.627

0.627
0.977

99% >0.99
1%

49
77%
15
23%
6.6  2.2

36 95%
2
5%
5.8  1.8

85
83% 0.026
17
17%
6.3  2.0
0.075

64

38

100% 102

100%

100%

15

SF-12 PCS was not jet observed (mean MCS in smokers 41.37
(95% CI of 36.76, 45.97) and in nonsmokers 45.40 (95% CI of 42.28,
48.52; p = 0.137); mean PCS in smokers 32.50 (95% CI of 30.16,
34.84) and in nonsmokers 36.21 (95% CI of 33.82, 38.61; p = 0.534).
Four weeks after surgery, the average VAS pain was 2.7 points in
smokers (mean; 95% CI of 2.04, 3.42) and 2.44 points in
nonsmokers (mean; 95% CI of 1.95, 2.93; p = 0.496) and had
therefore decreased by 3.8 points in smokers (95% CI of 4.90,
2.62) and by 3.3 points in nonsmokers (95% CI of 3.98, 2.62;
p = 0.458). Both smokers and nonsmokers reported increase in the
HRQoL as compared to preoperative values. Here, the MCS (mean
47.46 (95% CI of 42.08, 52.84) in smokers and 51.87 (95% CI of 49.14,
54.59) in nonsmokers, p = 0.103) increased more than the PCS
(mean 35.01 (95% CI of 32.49, 37.52) in smokers and 36.21 (95% CI
of 33.82, 38.61) in nonsmokers, p = 0.534).
One year after surgery, average VAS pain levels were recorded
as 3.36  3.2 (mean  SD) and maximal VAS levels were 3.94  3.87
in smokers. In nonsmokers, mean VAS values were 1.95  3.7 for
average (p = 0.078) and 2.72  5.1 for maximal pain levels
(p = 0.105). The overall group average SF-12 MCS score was 52.03
(95% CI of 49.71, 54.34). The respective overall group SF-12 PCS
score was 44.32 (95% CI of 41.65, 46.99). When stratied for
smoking, a signicant difference in the SF-12 MCS (smokers: mean
47.29 (95% CI of 41.64, 52.93); nonsmokers: mean 53.90 (95% CI of
51.63, 56.17); p = 0.0091) was found with equal SF-12 PCS scores
between the groups (smokers: mean 43.95 (95% CI of 38.49, 49.42);
nonsmokers: mean 44.46 (95% CI of 41.29, 47.62); p = 0.866).
3.4. Analysis of the primary outcome
From a univariate and multivariate perspective, smoking status
did not predict responder status of the SF-12 MCS and SF-12 PCS
(Tables 3 and 4): on the MCS metric (Table 3), smokers were 111%
(OR = 1.11) as likely as nonsmokers to achieve a favourable HRQoL
response to LDH surgery (p = 0.86). We are 95% condent that this
value lies between 0.35 and 3.51, indicating that nonsmoking
patients when compared to smoking patients are not more likely to

Fig. 1. Observed vs. imputed histograms. Demonstrates that the imputed variables resemble the observed, although with more variance.

16

M.N. Stienen et al. / Clinical Neurology and Neurosurgery 122 (2014) 1219

Table 3
Mental component score (MCS): univariate and multivariate logistic regression estimates of predictors of responsiveness using 35 multiply imputed datasets.
Univariate

Multivariate

Total imputation
variance

% increase in SE

Nonsmokers
Smokers
Age

OR
1.00
1.02
0.99

p-value

0.97
0.70

95% CI

0.37
0.96

2.84
1.02

OR
1.00
1.11
0.97

p-value

0.86
0.23

95% CI

0.35
0.92

1.00
3.51
1.02

0.34

24.8

0.0007

25.2

Sex
Male
Female

1.00
2.24

0.14

0.77

6.56

1.00
2.29

0.16

0.72

7.32

0.35

20.33

Previous lumbar surgery

Yes
1.00
No
1.46

0.49

0.50

4.24

1.00
1.45

0.55

0.42

5.05

0.40

25.8

Tertiary education
1.00
Yes
No
2.14

0.21

0.65

7.03

1.00
2.19

0.27

0.53

8.95

0.51

28.2

Disability payments
No
1.00
Yes
1.21

0.80

0.27

5.33

1.00
1.63

0.59

0.27

9.79

0.83

18.9

Currently working
Yes
1.00
No
2.05

0.26

0.59

7.11

1.00
5.2

0.72

0.86

32.39

0.85

14.6

of surgery and length of postoperative hospitalization did not

differ (Table 2).

achieve a favourable result as measured by the SF-12 MCS

responder grouping. On the PCS metric (Table 4), smokers were
118% (OR = 1.18) as likely as nonsmokers to achieve a favourable
HRQoL response to LDH surgery (p = 0.79). We are 95% condent
that this value lies between 0.35 and 3.98, indicating that
nonsmoking patients, when compared to smoking patients, are
not more likely to achieve a favourable result as measured by the
SF-12 PCS responder grouping.

3.6. Discharge
Interestingly, we found that signicantly more nonsmokers
(23%) were released into a stationary rehabilitation program than
smokers (5%, p = 0.026). This related to a slightly longer hospitalization time in nonsmokers by trend (mean 6.6 (nonsmokers) vs.
5.8 days (smokers, p = 0.075)), as in Switzerland it is not allowed for
patients to leave the clinic 24 h if a stationary rehabilitation is
planned after the acute hospitalization. In contrast to the
US-approach where lumbar disc surgery is sometimes done in
an outpatient setting, at our institution it is generally

3.5. Intra- and postoperative complications

Perioperative complication rates were similarly low without a
difference between smokers and nonsmokers. There were no
severe complications recorded in either group. Also, EBL, length

Table 4
Physical component summary (PCS): univariate and multivariate logistic regression estimates of predictors of responsiveness using 35 multiply imputed datasets.
Univariate

Multivariate

Total imputation
variance

% increase in SE

Nonsmokers
Smokers
Age

OR
1.00
1.00
0.98

p-value

0.99
0.50

95% CI

0.33
0.94

2.99
1.03

OR
1.00
1.18
0.96

p-value

0.79
0.16

95% CI

0.35
0.91

3.98
1.01

0.38

18.9

0.0008

19.4

Sex
Male
Female

1.00
1.79

0.32

0.56

5.72

1.00
2.58

0.14

0.74

9.00

0.40

14.6

Previous lumbar surgery

Yes
1.00
No
1.36

0.60

0.43

4.32

1.00
1.11

0.88

0.30

4.10

0.44

19.9

Tertiary education
Yes
1.00
No
0.35

0.21

0.07

1.84

1.00
0.26

0.15

0.04

1.65

0.89

18.3

Disability payments
No
1.00
Yes
0.88

0.89

0.15

5.15

1.00
0.74

0.77

0.10

5.65

1.06

26.4

Currently working
Yes
1.00
No
1.92

0.41

0.40

9.19

1.00
4.54

0.126

0.65

31.54

0.97

10.9

M.N. Stienen et al. / Clinical Neurology and Neurosurgery 122 (2014) 1219

recommended to patients to stay at least until day 2 or 3 after

surgery in order to establish regular wound care, adapt analgesia
and receive in-house physiotherapy. Elderly patients are usually
discharged on days 57. Swiss stationary rehabilitation programs
aim to improve lost or misdirected functions and require an active
participation of the patient. Patients receive intensive daily
physiotherapy over a course of usually 23 weeks to improve
mobilization, stabilization and coordination by means of modern
and effective techniques. In addition, pain medication is adapted
on daily basis, if necessary, and patients are assisted in social
regards like return to work or occupational retraining.
4. Discussion
This study prospectively assessed the VAS pain and HRQoL
outcomes as well as complication rates of smokers and nonsmokers with LDH requiring surgery. Overall, patients of both
study groups proted equally from surgery and no difference was
found between complication rates and the outcomes of smokers
and nonsmokers. Although it is well known that patients who
smoke have poorer outcomes on a wide variety of procedures and
metrics, we conclude on the basis of our data that if an effect of
smoking on outcomes after lumbar disc surgery existed, the effect
size would be rather small. Therefore, there appears to be no
reason to alter the short-term smoking habits, and instead
maintain focus on the long-term smoking cessation.
Nevertheless, smoking seems to affect the lumbar spine as
Helivaara et al. demonstrated a higher incidence of low back pain
(LBP) in patients smoking more than 20 cigarettes daily in their
analysis of 5500 patients OR = 1.9 (male) and 2.7 (female)) [25].
Likewise, Kelsey et al. reported an association between symptomatic LDH and last years smoking status (OR = 1.7). The risk
increased by 20% per every 10 cigarettes smoked daily for the
last year. In contrast, former smokers with smoking cessation for
1 year did not reveal an increased risk for LDH (OR = 1.0) [16].
Despite the previous literature on clinical studies indicating a
robust association between smoking and LBP or disc disease, the
exact pathophysiological mechanisms have not been fully
claried. It has proved difcult to establish a direct pathophysiological correlation between LBP and smoking, as the impact of
tobacco use on the back muscles and the axial skeleton is
thought to be multifactorial [2]. Tobacco use is a major risk
factor for cardiovascular disease and peripheral arterial circulatory disorders. Along with other cardiovascular risk factors
(cvRF), it increases the likelihood of developing spinal degeneration due to malnutrition and hypoxemia [26,27]. Reasons for
this include reduced hemoglobin and tissue oxygenation [28],
nicotine-triggered vasoconstriction [29] and arteriosclerosis
[30], and hematorheological factors [31] including increase of
plasma brinogen, reduced plasminogen activator and sensitization of blood platelets leading to increased viscosity and
general increase of the coagulability [32,33]. The effects of
smoking on the intervertebral discs were studied by Holm and
Nachemson, who in particular regarded the above-mentioned
cvRFs and increased carbon monoxide blood levels responsible
for degeneration [27]. Moreover, nicotine itself seems to play a
role as both low (100 mg/mL [34]), and very high doses of
nicotine (800050,000 mg/mL [26]) inhibited nucleus pulposus
cell metabolism and the production of glycosaminoglycans and
collagen in vitro [35]. Other theories comprise inammatory
mechanisms with possible excess immune responses against
discal glycoproteins in smokers [36,37].
This study was balanced for all potential confounders except for
a higher incidence of sensory decits that was recorded in the
group of smokers. As no difference in the baseline functional
impairment was noted between the two groups, we believe that

17

this difference did not bias our results. A likewise higher incidence
of sensory decits in smoking patients after radial artery
harvesting for coronary bypass surgery has been reported without
inuence on the long-time morbidity [38]. Smokers presented
with a tendency towards worse mean MCS values at baseline
(mean difference, 4.39; 95% CI, .489.28), a phenomenon that has
been addressed before [19,39]. Even though smokers started from
lower baseline MCS values, both study groups revealed similar
changes in the MCS values after surgery for LDH. We also regard the
observed trend towards a higher mean VAS at baseline in smoking
patients as clinically insignicant (mean difference, 0.860) [40,41].
As outlined in Table 2, similar surgical procedures were performed
in smokers and nonsmokers, thus not biasing our results.
The fact that a previously smoking and a never smoking
patient sample was combined and compared to a currently
smoking cohort carries some risk, taking into consideration that
the lifetime burden of cigarette smoking could inuence the
primary outcome [4244]. However, some studies suggest that
the smoking status at time of surgery may be more important
than the absolute number of cigarettes smoked. As such, Kelsey
et al. found the OR for a prolapsed lumbar disc of both neversmokers and previous smokers (1 year smoking cessation) to be
1.0, while current smokers at time of surgery had an OR of 1.7
(95% CI, 1.02.5) in a case-control study with age- and gendermatched controls [16]. In addition, previous reports were able to
demonstrate that perioperative smoking cessation can be helpful
in reversing the impact of smoking on postoperative outcomes
[9,45]. Thus, the negative effect of smoking may be reversible
within weeks through months. In our study, we aimed to study,
whether the smoking status at time of surgery negatively
inuenced the outcome of our patients. The smoking status at
time of surgery seemed to be more adequate for our study
purpose than the lifetime amount of cigarettes consumed and
was therefore taken to build our patient groups.
Previous studies on smoking and outcomes after spinal surgery
have handled this point differently, while most studies relied on
the same grouping mechanism as we did. Andersen et al. found the
negative effects of bony fusion after lumbar fusion surgery most
pronounced in patients who smoked at the time of surgery as
compared with never-smokers and previous smokers. They found
smoking predicted nonunion (OR, 2.01; 95% CI, 1.143.55, p = 0.016)
and patient dissatisfaction with the operation result (OR, 1.65; 95%
CI, 0.992.74, p = 0.054) [12]. Glassman et al. identied nonunion in
24 (14.2%) of 169 nonsmokers compared with 39 (20.7%) of 188
smokers they also used time of surgery as reference point for
study grouping [9]. However, the authors went one step further
and additionally looked at nonunion rates in patients with low vs.
heavy smoking habit and patients who quit smoking either before
or for a variable period after surgery. The nonunion rate for those
patients who were unable to quit smoking before surgery was
22.2% compared with 19.6% for patients who quit for longer than
1 month before surgery. Postoperative smoking cessation altered
nonunion rates even more markedly. There was a 26.5% nonunion
rate (18/68) in patients who did not quit smoking after surgery as
opposed to 18.2% (8/44) in patients who quit smoking for 16
months after surgery and 17.1% in patients who quit for >6 months.
Despite the intuitive perception that risk for nonunion should be
incrementally linked to the quantity that a patient smokes, the
data did not support this assumption. Nonunion rates were 23.5%,
19.8%, and 19.3% for patients smoking less than 1 pack per day
(ppd), 12 ppd and >2 ppd, respectively. Therefore, the smoking
status during the year of surgery seems to be more relevant for
spinal fusion than the absolute amount of cigarettes consumed,
also supporting our approach used here to compare patient groups
in respect of their smoking status at time of surgery. Likewise,
Betagnoli et al. used smoking status at time of surgery as reference

18

M.N. Stienen et al. / Clinical Neurology and Neurosurgery 122 (2014) 1219

point to build their study groups. They found that up to 24 months

after lumbar total disc arthroplasty utilizing the ProDisc prothesis
postoperative outcomes between smoking and nonsmoking
patients were similar [13]. An et al. compared smoking habits
between 205 patients with lumbar and cervical disc disease at time
of surgery and 205 age- and sex-matched inpatient controls. They
also discriminated between current smokers, former smokers and
never smokers. Patients with lumbar (56% vs. 37.5%, p = 0.0025)
and cervical disc disease (64.3% vs. 37.5%, p = 0.0025) were much
more likely to smoke than controls. The authors found that for both
lumbar and cervical disc disease the effect was more pronounced
when current smoking as opposed to previous smoking was used
for analysis [11]. Lee et al. looked at predictors for outcome after
conventional open lumbar discectomy in 40 patients. They
compared a smoking and nonsmoking cohort at time of surgery
and found no signicant difference in the Oswestry Disability
Index at 1-year postoperative [17]. One of the largest studies on the
effect of smoking on outcomes after lumbar spine surgery (n = 4555
patients from the Swedish Spine Register) used the smoking status
at time of surgery to built study groups and demonstrated that
smokers proted signicantly less from decompression surgery for
lumbar spinal stenosis [19]. There are numerous further studies in
the literature using the time of surgery as reference point to build
their study groups, some of which also combined previous smokers
and nonsmokers in their analysis (as we did here) [14,15,18,2022].
In contrast, a study using national survey data by Deyo et al.
found that low back pain may be sensitive to the cumulative
patient exposure to cigarette smoke [46]. This was not a study
comparing post-surgical outcomes, but it still indicates that there
may also be a dose-related effect of smoking on low back pain that
should be considered when interpreting our results. Our smoking
collective comprised both heavy and light smokers, which could be
a possible reason for not nding a signicant effect. Appaduray
et al. retrospectively analyzed complications after spinal surgery in
902 patients and found positive smoking history (without
discriminating if patients smoked at time of surgery) in the patient
chart not to be associated with a higher complication rate [47].
When compared to a German reference collective with LBP that
was managed conservatively with rehabilitation (n = 169; PCS:
35.07  9.69; MCS: 47.51 10.72), our patients scored generally
worse on the mean baseline PCS and MCS by about 5 points each
(Table 1) [10]. These baseline differences indicate that our patients
with lumboradicular pain were in a worse condition, which is also
reected by the fact that surgery had to be performed in all
patients of our cohort. When managed conservatively, the PCS was
40.20  9.94 and MCS was 49.22  10.02 (n = 169) after rehabilitation, referring to an increase in the PCS and MCS of 5 and 2 points,
respectively [10]. Our results show an increase in the mean PCS and
MCS of 1.66 and 1.13 points immediately after surgery, of 4.66 and
7.92 4 weeks after surgery and of 14.1 and 9.16 at the 1-year followup. Thus, irrespective of the smoking status, the surgical therapy
proved effective and stable MCS and PCS values could be obtained,
correlating to the patients actual condition. One year after surgery,
both mean MCS and PCS values revealed a trend towards the
German-speaking population norm (n = 2805; PCS 49.0  9.4; MCS
52.2  8.1) [10].
Interestingly, a signicant improvement in the PCS was
recorded already few days after surgery in the combined groups
(mean change 1.66, p < 0.001), while the mental status improved
less at this point (mean change 1.13, p < 0.001). We gure, that
while the patients mobility increases immediately after surgery,
their sense of ill-being is not signicantly affected until discharge
from the hospital. As often noticed and discussed by clinicians of
all surgical subspecialties, we also observed that the smoking
participants of our study were more likely to mobilize faster due to
their wish to leave the ward to smoke outside. Still, this

phenomenon was not recorded for study purposes and its effect
on the outcome could thus not be analyzed. After 4 weeks, the MCS
increased more than the PCS with only little more increase at the
1-year follow-up. This reects the clinical observation that
patients after lumbar surgery are still physically restricted with
good subjective well being at the follow-up consultation 4 weeks
after surgery. We interpret the fact that a signicant higher
fraction of nonsmokers was released into a stationary rehabilitation program with a generally higher health consciousness in
nonsmoking patients. However, this hypothesis cannot be proven
by our data.
Our study was not designed for a longer follow-up than 1 year
postoperative and therefore our data are insufcient to show
possible longer-lasting adverse effects of smoking. There is a trend
in evaluation of long-term outcomes after spine surgery including
pain, function and HRQoL measures. The prospective, randomized
SPORT study had been designed to compare 2-year ODI, SF-36, and
further outcomes after surgical or best medical therapy for patients
with intervertebral disc herniation and persistent symptoms
despite some non-operative treatment for at least 6 weeks [48].
The follow-up was continued and 4-year outcomes have been
published [49]. Likewise, the follow-up of the prospective
randomized study by the Leiden-The Hague Spine Intervention
Prognostic Study Group, using Roland-Disability-Questionnaire
and VAS pain levels amongst others, has been prolonged for a
5-year outcome, [50] after 1-year results had been published [51].
Previous prospective studies evaluating if smoking affected
outcomes mostly used 1-year [15,17,21] follow-up periods after
surgery for lumbar disc herniation, while one study collected
outcomes after 2 years and a mean of 7.3  1 years [20]. One-year
[52] or 2-year [19] follow-up periods after lumbar decompression
surgery, and 1-year [14] or 2-year [12,13,18] follow-up periods after
fusion surgery have been reported. While bony fusion happens
over a longer interval and therefore must be followed longer, most
of the effect of lumbar discectomy on pain-, functional- and
HRQoL-outcomes is exerted within 12 months after the index
procedure [20,4851]. Still, a long-term follow up would have
given more evidence concerning possible adverse effects of
smoking on the primary outcome.
The high rate of loss of follow-up was probably due to the
method of a mailed questionnaire and could have been prevented
by using a standardized telephone interview. The use of multiple
imputation minimized the missing data effect on the primary
outcome variable. Therefore, the missing data mechanism has
been handled by the use of multiple imputation and subsequently
our results are congruent with other studies on HRQoL after
lumbar spine surgery.
5. Conclusions
Our hypothesis that smoking patients suffered from more and
longer pain and displayed worse HRQoL scores until 1 year after
surgery for LDH had to be rejected as results were comparable to
those of nonsmoking patients. If an effect of smoking on these
variables existed, the effect size would be small. Further prospective cohort studies are mandatory to clarify the inuence of the
smoking status on pain levels and HRQoL after surgery for LDH.
Acknowledgement
The authors thank all patients who took part in the study.
References
[1] Turan A, Mascha EJ, Roberman D, Turner PL, You J, Kurz A, Sessler DI, Saager L.
Smoking and perioperative outcomes. Anesthesiology 2011;114:83746.

M.N. Stienen et al. / Clinical Neurology and Neurosurgery 122 (2014) 1219
[2] Stienen MN, Richter H, Prochnow N, Schnakenburg LF, Gautschi OP. [Does
smoking correlate with low back pain and the outcome of spinal surgery?]. Z
Orthop Unfall 2011;149:31723.
[3] Myles PS, Iacono GA, Hunt JO, Fletcher H, Morris J, McIlroy D, Fritschi L. Risk of
respiratory complications and wound infection in patients undergoing
ambulatory surgery: smokers versus nonsmokers. Anesthesiology
2002;97:8427.
[4] Wong J, Lam DP, Abrishami A, Chan MT, Chung F. Short-term preoperative
smoking cessation and postoperative complications: a systematic review and
meta-analysis. Can J Anaesth 2012;59:26879.
[5] Sorensen LT. Wound healing and infection in surgery. The clinical impact of
smoking and smoking cessation: a systematic review and meta-analysis. Arch
Surg 2012;147:37383.
[6] Thomsen T, Tonnesen H, Moller AM. Effect of preoperative smoking cessation
interventions on postoperative complications and smoking cessation. Br J Surg
2009;96:45161.
[7] Brown CW, Orme TJ, Richardson HD. The rate of pseudarthrosis (surgical
nonunion) in patients who are smokers and patients who are nonsmokers: a
comparison study. Spine (Phila Pa 1976) 1986;11:9423.
[8] Wing KJ, Fisher CG, OConnell JX, Wing PC. Stopping nicotine exposure before
surgery. The effect on spinal fusion in a rabbit model. Spine (Phila Pa 1976)
2000;25:304.
[9] Glassman SD, Anagnost SC, Parker A, Burke D, Johnson JR, Dimar JR. The effect
of cigarette smoking and smoking cessation on spinal fusion. Spine (Phila Pa
1976) 2000;25:260815.
[10] Morfeld M, Kirchberger I, Bullinger M. SF-36.Fragebogen zum Gesundheitszustand. Deutsche Version des Short Form-36 Health Survey. 2. Prague, Toronto,
Cambridge (MA), Amsterdam, Kopenhagen, Stockholm, Gttingern Bern,
Wien, Paris, Oxford: Hogrefe Verlag; 2011 ergnzte und berarbeitete Auage.
[11] An HS, Silveri CP, Simpson JM, File P, Simmons C, Simeone FA, Balderston RA.
Comparison of smoking habits between patients with surgically conrmed
herniated lumbar and cervical disc disease and controls. J Spinal Disord
1994;7:36973.
[12] Andersen T, Christensen FB, Laursen M, Hoy K, Hansen ES, Bunger C. Smoking
as a predictor of negative outcome in lumbar spinal fusion. Spine (Phila Pa
1976) 2001;26:26238.
[13] Bertagnoli R, Yue JJ, Kershaw T, Shah RV, Pfeiffer F, Fenk-Mayer A, Nanieva R,
Karg A, Husted DS, Emerson JW. Lumbar total disc arthroplasty utilizing the
ProDisc prosthesis in smokers versus nonsmokers: a prospective study with 2year minimum follow-up. Spine (Phila Pa 1976) 2006;31:9927.
[14] Cobo Soriano J, Sendino Revuelta M, Fabregate Fuente M, Cimarra Diaz I,
Martinez Urena P, Deglane Meneses R. Predictors of outcome after
decompressive lumbar surgery and instrumented posterolateral fusion. Eur
Spine J 2010;19:18418.
[15] Jansson KA, Nemeth G, Granath F, Jonsson B, Blomqvist P. Health-related
quality of life in patients before and after surgery for a herniated lumbar disc.
J Bone Joint Surg Br 2005;87:95964.
[16] Kelsey JL, Githens PB, OConner T, Weil U, Calogero JA, Holford TR, White AA.
Acute prolapsed lumbar intervertebral disc. An epidemiologic study with
special reference to driving automobiles and cigarette smoking. Spine (Phila Pa
1976) 1984;9:60813.
[17] Lee JC, Kim MS, Shin BJ. An analysis of the prognostic factors affecing the
clinical outcomes of conventional lumbar open discectomy: clinical and
radiological prognostic factors. Asian Spine J 2010;4:2331.
[18] Luca A, Mannion AF, Grob D. Should smoking habit dictate the fusion
technique?. Eur Spine J 2011;20:62934.
[19] Sanden B, Forsth P, Michaelsson K. Smokers show less improvement than
nonsmokers two years after surgery for lumbar spinal stenosis: a study of 4555
patients from the Swedish spine register. Spine (Phila Pa 1976) 2011;36:105964.
[20] Silverplats K, Lind B, Zoega B, Halldin K, Rutberg L, Gellerstedt M, Brisby H.
Clinical factors of importance for outcome after lumbar disc herniation
surgery: long-term follow-up. Eur Spine J 2010;19:145967.
[21] Slover J, Abdu WA, Hanscom B, Weinstein JN. The impact of comorbidities on
the change in short-form 36 and oswestry scores following lumbar spine
surgery. Spine (Phila Pa 1976) 2006;31:197480.
[22] Smith JS, Shaffrey CI, Glassman SD, Carreon LY, Schwab FJ, Lafage V, Arlet V,
Fu KM, Bridwell KH. Spinal Deformity Study G. Clinical and radiographic
parameters that distinguish between the best and worst outcomes of scoliosis
surgery for adults. Eur Spine J 2013;22:40210.
[23] Luo X, Lynn George M, Kakouras I, Edwards CL, Pietrobon R, Richardson W, Hey
L. Reliability, validity, and responsiveness of the short form 12-item survey
(SF-12) in patients with back pain. Spine (Phila Pa 1976) 2003;28:173945.
[24] Mundry R, Nunn CL. Stepwise model tting and statistical inference: turning
noise into signal pollution. Am Nat 2009;173:11923.
[25] Heliovaara M, Makela M, Knekt P, Impivaara O, Aromaa A. Determinants of
sciatica and low-back pain. Spine (Phila Pa 1976) 1991;16:60814.
[26] Akmal M, Kesani A, Anand B, Singh A, Wiseman M, Goodship A. Effect of
nicotine on spinal disc cells: a cellular mechanism for disc degeneration. Spine
(Phila Pa 1976) 2004;29:56875.

19

[27] Holm S, Nachemson A. Nutrition of the intervertebral disc: acute effects of

cigarette smoking. An experimental animal study. Ups J Med Sci 1988;93:919.
[28] Jensen JA, Goodson WH, Hopf HW, Hunt TK. Cigarette smoking decreases
tissue oxygen. Arch Surg 1991;126:11314.
[29] Bruce J, Miller J, Hooker D. The effect of smoking upon the blood pressure and
upon the volume of the hand. Am J Physiol 1909;24:104.
[30] Pittilo RM, Mackie IJ, Rowles PM, Machin SJ, Woolf N. Effects of cigarette
smoking on the ultrastructure of rat thoracic aorta and its ability to produce
prostacyclin. Thromb Haemost 1982;48:1736.
[31] Ingvarsson RF, Bjarnason AO, Dagbjartsson A, Hardardottir H, Haraldsson A,
Thorkelsson T. The effects of smoking in pregnancy on factors inuencing fetal
growth. Acta Paediatr 2007;96:3836.
[32] Renaud S, Blache D, Dumont E, Thevenon C, Wissendanger T. Platelet function
after cigarette smoking in relation to nicotine and carbon monoxide. Clin
Pharmacol Ther 1984;36:38995.
[33] Rothwell M, Rampling MW, Cholerton S, Sever PS. Haemorheological changes
in the very short term after abstention from tobacco by cigarette smokers. Br J
Haematol 1991;79:5003.
[34] Nemoto Y, Matsuzaki H, Tokuhasi Y, Okawa A, Uematu Y, Nishimura T, Oda H.
Histological changes in intervertebral discs after smoking and cessation:
experimental study using a rat passive smoking model. J Orthop Sci
2006;11:1917.
[35] Kim KS, Yoon ST, Park JS, Li J, Park MS, Hutton WC. Inhibition of proteoglycan
and type II collagen synthesis of disc nucleus cells by nicotine. J Neurosurg
2003;99:2917.
[36] Noble RC, Penny BB. Comparison of leukocyte count and function in smoking
and nonsmoking young men. Infect Immun 1975;12:5505.
[37] Silverman NA, Potvin C, Alexander Jr. JC, Chretien PB. In vitro lymphocyte
reactivity and T-cell levels in chronic cigarette smokers. Clin Exp Immunol
1975;22:28592.
[38] Shah SA, Chark D, Williams J, Hessheimer A, Huh J, Wu YC, Chang PA, Scholl FG,
Drinkwater DC. Retrospective analysis of local sensorimotor decits after
radial artery harvesting for coronary artery bypass grafting. J Surg Res
2007;139:2038.
[39] Woolf SH, Rothemich SF, Johnson RE, Marsland DW. Is cigarette smoking
associated with impaired physical and mental functional status? An ofcebased survey of primary care patients. Am J Prev Med 1999;17:1347.
[40] Bird SB, Dickson EW. Clinically signicant changes in pain along the visual
analog scale. Ann Emerg Med 2001;38:63943.
[41] Gallagher EJ, Liebman M, Bijur PE. Prospective validation of clinically
important changes in pain severity measured on a visual analog scale. Ann
Emerg Med 2001;38:6338.
[42] Cancer IIAfRo. Reversal of Risk after Quitting Smoking, Vol. 11. WHO press; 2007.
[43] Doll R, Peto R, Boreham J, Sutherland I. Mortality in relation to smoking: 50
years observations on male British doctors. BMJ 2004;328:1519.
[44] Mahmud A, Feely J. Effect of smoking on arterial stiffness and pulse pressure
amplication. Hypertension 2003;41:1837.
[45] Mills E, Eyawo O, Lockhart I, Kelly S, Wu P, Ebbert JO. Smoking cessation
reduces postoperative complications: a systematic review and meta-analysis.
Am J Med 2011;124:14454 e148.
[46] Deyo RA, Bass JE. Lifestyle and low-back pain. The inuence of smoking and
obesity. Spine (Phila Pa 1976) 1989;14:5016.
[47] Appaduray SP, Lo P. Effects of diabetes and smoking on lumbar spinal surgery
outcomes. J Clin Neurosci 2013;20:17137.
[48] Weinstein JN, Lurie JD, Tosteson TD, Skinner JS, Hanscom B, Tosteson AN,
Herkowitz H, Fischgrund J, Cammisa FP, Albert T, Deyo RA. Surgical vs
nonoperative treatment for lumbar disk herniation: the spine patient
outcomes
research
trial
(SPORT)
observational
cohort.
JAMA
2006;296:24519.
[49] Weinstein JN, Lurie JD, Tosteson TD, Tosteson AN, Blood EA, Abdu WA,
Herkowitz H, Hilibrand A, Albert T, Fischgrund J. Surgical versus nonoperative treatment for lumbar disc herniation: four-year results for the Spine
Patient Outcomes Research Trial (SPORT). Spine (Phila Pa 1976)
2008;33:2789800.
[50] Lequin MB, Verbaan D, Jacobs WC, Brand R, Bouma GJ, Vandertop WP, Peul
WC. Leiden The Hague spine intervention prognostic study G, Wcp,
Bwk, Rtwmt, WbvdH, Rb. Surgery versus prolonged conservative
treatment for sciatica: 5-year results of a randomised controlled trial.
BMJ Open 2013;3.
[51] Peul WC, van Houwelingen HC, van den Hout WB, Brand R, Eekhof JA, Tans JT,
Thomeer RT, Koes BW. Leiden The Hague spine intervention prognostic study
G. Surgery versus prolonged conservative treatment for sciatica. N Eng J Med
2007;356:224556.
[52] Jansson KA, Nemeth G, Granath F, Jonsson B, Blomqvist P. Health-related
quality of life (EQ-5D) before and one year after surgery for lumbar spinal
stenosis. J Bone Joint Surg Br 2009;91:2106.

Reproduced with permission of the copyright owner. Further reproduction prohibited without
permission.