SEARCH

SEARCH BY CITATION

Keywords:

  • Smoking: anaesthesia, postoperative complications, intensive care admission

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Study population
  5. Study procedure
  6. Statistical analysis
  7. Results
  8. Discussion
  9. Acknowledgements
  10. References

Summary Smoking is a risk factor for intra-operative pulmonary complications and a wide range of postoperative pulmonary, cardiovascular, infection and wound-related complications. These may all lead to unplanned postoperative intensive care admission. We tested the hypothesis that smokers have an increased incidence of postoperative intensive care admission and more postoperative complications than nonsmokers in a general and orthopaedic surgical population. The following information was assessed in 6026 surgical patients: age, sex and smoking status (pack-years), history of heart and lung disease, American Society of Anesthesiologists (ASA) physical classification, intensive care admission and postoperative complications. Two thousand five hundred and twenty-six (46%) were smokers but for 620 patients (10.3%) smoking status was not confirmed. Postoperative intensive care admission was required by 319 patients (5.3%). Patients with > 50 pack-years were admitted to the intensive care more frequently than were smokers with ≤50 pack-years history and nonsmokers (p < 0.001). Ex-smokers with > 50 pack-years history had the same risk of postoperative admission to intensive care as smokers with > 50 pack-years history. Smokers admitted to intensive care with > 50 pack-years history had a higher incidence of chronic lung disease (p < 0.005) and heavy alcohol consumption (p < 0.001). These smokers also had a higher incidence of postoperative pulmonary complications (odds ratio = 3.91, p < 0.01). The mortality rate was 37% in smokers with > 50 pack-years history and 24% in nonsmokers (odds ratio = 2.02, p = 0.08). We conclude long-term tobacco smoking (> 50 pack-years) carries a higher risk of postoperative admission to intensive care, and there seems to be a dose relationship between the amount of tobacco consumed and the risk of postoperative intensive care admission.

Smoking is an important risk factor for a wide range of postoperative complications, especially pulmonary and cardiovascular complications [1–6]. Smoking causes reduced closing capacity, poorer clearing of pulmonary secretions and chronic obstructive lung disease [3, 5], as well as impaired collagen synthesis [1, 2] and impaired cardiovascular [6] and immune functions [7, 8]. We have previously described the risk of tobacco smoking in relation to unplanned intensive care admission [9], but have not focused on the effect of long-term smoking on postoperative morbidity. The aim of this study was to investigate the dose relationship between smoking and postoperative intensive care admission and complications in a general and orthopaedic surgical population.

Study population

  1. Top of page
  2. Abstract
  3. Methods
  4. Study population
  5. Study procedure
  6. Statistical analysis
  7. Results
  8. Discussion
  9. Acknowledgements
  10. References

We included patients scheduled for general or orthopaedic surgery between 1 May 1998 and 30 April 2000 at Bispebjerg University Hospital, Copenhagen. The local ethics committee approved the study. Smoking status was recorded as pack-years, i.e. (daily consumption × number of years smoking)/20.

Study procedure

  1. Top of page
  2. Abstract
  3. Methods
  4. Study population
  5. Study procedure
  6. Statistical analysis
  7. Results
  8. Discussion
  9. Acknowledgements
  10. References

A special record sheet was designed for the study. This included descriptive information about the patient [age, sex, smoking status (age when started smoking, daily consumption, age when stopped smoking if an ex-smoker)], alcohol consumption (units per week), history of chronic pulmonary or cardiac disease, American Society of Anaesthesiologists' (ASA) physical classification, information about type of surgery and anaesthesia, causes of postoperative admission to intensive care unit (ICU) and any complication(s).

A study team of three doctors conducted the search for, and reviewed the charts of, those patients admitted to ICU. The pulmonary complications recorded included pneumonia (fever, chest X-ray changes or positive sputum culture), atelectasis seen on chest X-ray, pneumothorax (diagnosed on chest X-ray), and respiratory insufficiency (requiring medical intervention). The cardiovascular complications recorded were congestive heart failure with bilateral rales, pulmonary oedema or left-sided heart failure documented by usual X-ray changes, cardiac arrest, heart block, second- or third-degree atrioventricular block or bundle branch block, ventricular tachyarrhythmias and acute myocardial infarction.

Mortality in ICU or from a complication related to the original procedure was recorded.

Statistical analysis

  1. Top of page
  2. Abstract
  3. Methods
  4. Study population
  5. Study procedure
  6. Statistical analysis
  7. Results
  8. Discussion
  9. Acknowledgements
  10. References

The admission rate to ICU was stratified according to pack-years in the intervals: = 10, 10–30, 31–50, 51–70 and > 70. To reduce the risk of multiple statistical testing, we used a significance level of 0.01. In the analysis of the 319 patients admitted to ICU, for nominal and/or categorical variables, we used the Chi-squared test and in cases where expected frequencies were small Fisher's exact test was used. For comparison of numeric variables, Mann–Whitney's U-test was used. Adjusted odds ratio interval was computed for the risk variables with 99% confidence intervals (CI). Statistical analyses were performed with the SAS® (SAS Institute, Cary, NC, USA).

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Study population
  5. Study procedure
  6. Statistical analysis
  7. Results
  8. Discussion
  9. Acknowledgements
  10. References

Six thousand and twenty-six patients entered the study. We did not study 508 patients whose charts were missing after repeated search, as well as 23 patients less than 15 years old, and 112 patients whose smoking status was unknown. Forty-six per cent (2526/5495) were smokers, and in 620 of the patients (10.3%) smoking status was unspecified or unknown (Fig. 1). Three hundred and nineteen patients (5.3%) required postoperative intensive care admission; their type of surgery and reasons for ICU admission are given in Tables 1 and 2. No statistically significant differences were found among the smokers and nonsmokers with respect to type of surgery and anaesthesia. The incidence of postoperative intensive care admission in smokers was 6.0% compared with 4.9% in nonsmokers with odds ratio of 1.23 (99% CI = 0.95−1.58, p = 0.14).

image

Figure 1. Trial profile.

Download figure to PowerPoint

Table 1.  Type of surgery in 319 patients admitted to ICU.
Acute major abdominal (explorative laparotomy,  perforated peptic ulcer, removal of adhesions
 in intestinal obstruction, nephrectomy)151
Gastrointestinal (cholecystectomy, laparoscopic  cholecystectomy, gastrojejunostomy, partial  and total colectomy, rectosigmoidectomy,  retroperitoneal tumorectomy)69
Minor abdominal (herniotomy, appendectomy)16
Acute major orthopaedic (reduction and internal  fixation of major fractures)38
Major orthopaedic surgery (hip arthroplasty)5
Other minor surgical procedures (gastroscopy,  cystoscopy,dilatation and curettage, laparoscopy)40
Table 2.  Reasons for postoperative intensive care admission in the 319 patients (a patient may have more than one cause).
Post-operative stabilisation (bronchodilators,  pharmacologic intervention due to evidence
 of myocardial ischaemia, arrthythmias or  new infiltrates on chest X-ray)108
Cardiovascular failure (circulatory failure with  systolic blood pressure < 90 mmHg, lung  oedema, peri-operative cardiac arrest or  suspected myocardial infarction)80
Pulmonary failure [sign of respiratory insufficiency  necessitating intervention with continuous  positive airway pressure (CPAP), bronchodilators,  re-intubation, or mechanical ventilation  or pneumonia requiring ventilation]66
Metabolic and/or renal failure (disordered fluid  balance, hypovolaemia, acid−base balance, sodium  and potassium disorders, creatinine > 0.3 mmol.l−1 or dialysis)74
Surgical complications (bleeding, haematoma,  wound related infection, sepsis)58

Table 3 shows the ICU admission rate for smokers and ex-smokers stratified according to pack-years.

Table 3.  Distribution of smokers and ex-smokers between ward and ICU stratified according to smoking history.
Pack-yearsSmokers + ICU n (%) Ward n (%) Ex-smokers + ICU n (%) Ward n
  1. +Indicates the maximum number of patients in the analysis; *p < 0.001 indicates a higher rate of smokers in ICU compared with smokers with pack-years = 10; **p < 0.001 indicates a higher rate of smokers in ICU compared to smokers with pack-years 31–50, 11–30 and ≤10; and ***p < 0.001 indicates a higher rate of ex-smokers in ICU compared to ex-smokers with pack-years 11–30 and ≤10.

≤1012 (2.0)6032 (1.4)148
11–3060 (6.1)*9248 (4.4)174
31–5030 (5.7)*4928 (11.3)63
51–7029 (14.1)**1755 (15.6)***27
> 7012 (13.5)**773 (11.1)24

Smokers with > 50 pack-years were more often admitted to ICU than smokers with ≤50 pack-years. Ex-smokers with > 50 pack-years had almost the same risk of postoperative ICU admission as smokers with > 50 pack-years. Smokers in ICU with > 50 pack-years were younger than nonsmokers (p < 0.05). They also had a higher incidence of chronic lung disease (p < 0.005) and a heavy alcohol consumption (exceeding 35 units/week, p < 0.001) (Table 4). Smokers with > 50 pack-years had a higher incidence of postoperative pulmonary complications during the stay in ICU with odds ratio of 3.91 (99% CI = 1.42–10.79, p < 0.01) (Table 4). We found no difference in postoperative cardiovascular complications between smokers and nonsmokers. The rate of postoperative death was 37% in smokers with > 50 pack-years compared with 24% in nonsmokers with odds ratio of 2.02 (99% CI = 0.92–4.41, p = 0.08).

Table 4.  Pre-operative and postoperative characteristics of the patient's admission to ICU.
CharacteristicsSmokersEx-smokersNonsmokers
> 50 pack year≤50 pack year > 50 pack year≤ 50 pack year
(n = 41)(n = 102)(n = 8)(n = 18)(n = 108)
  • a

    p < 0.05: indicates a higher median age in nonsmokers and ex-smokers compared to smokers;

  • b

    p < 0.005: indicates a higher rate of chronic lung disease in smokers with pack-years > 50 than smokers with ≤50 pack-years and nonsmokers;

  • c

    p < 0.0001: indicates a higher rate of daily alcohol consumption in smokers than nonsmokers;

  • d

    p < 0.01: indicates a higher rate of daily alcohol consumption in ex-smokers than nonsmokers;

  • e

    p < 0.01: indicates a higher rate of smokers with pack-years > 50 than nonsmokers;

  • f

    p = 0.08: comparison between smokers with pack-years > 50 years and nonsmokers.

Preoperative
Age; year median (range)73 (41–88)71 (26–87)79a (66–86)77a (51–86)77a (37–94)
Sex (women); n (%)23 (56)48 (47) 6 (75) 7 (39)68 (63)
Chronic lung disease; n (%)22 (54)b28 (28) 4 (50) 5 (28)23 (21)
Chronic heart disease; n (%)26 (63)40 (39) 7 (88)13 (72)52 (48)
ASA-classification
 I−II; n (%)12 (29)38 (37) 4 (50) 6 (33)36 (33)
 III−V; n (%)29 (71)64 (63) 4 (50)12 (67)72 (67)
Alcohol consumption:
 Daily; n (%)32 (78)c76 (75)c 7 (88)d14 (78)d45 (42)
 ≤17 units/week; n (%)19 (46)61 (60) 6 (75)10 (56)38 (35)
 18–34 units/week; n (%)5 (12) 5 (5) 0 2 (11) 4 (4)
 ≤35 units/week; n (%)8 (20)10 (8) 1 (13) 2 (11) 3 (3)
Post-operative
Pulmonary complications; n (%)36 (88)e77 (76) 8 (100)14 (78)70 (65)
Cardiovascular complications; n (%)25 (61)53 (52) 3 (38)13 (72)57 (53)
Death; n (%)15 (37)f26 (26) 2 (25) 6 (33)24 (22)

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Study population
  5. Study procedure
  6. Statistical analysis
  7. Results
  8. Discussion
  9. Acknowledgements
  10. References

In this study, we found a strong association between long-term tobacco smoking and postoperative ICU admission. Smokers with > 50 pack-years had an increased risk of ICU admission after surgery, compared with smokers with ≤50 pack-years. A dose relationship between pack-years of cigarette smoking and risk of postoperative ICU admission was found. Smoking > 50 pack-years is an important risk factor for serious postoperative complications. This is probably due to chronic changes in pulmonary [3–5], cardiovascular [6] and immune function [2, 7, 8]. These chronic changes may well be the reason that ex-smokers who have smoked > 50 pack-years have the same risk of intensive care admission as the smokers who have smoked > 50 pack-years.

A noticeable factor is that the smokers admitted to ICU postoperatively are younger than their nonsmoking counterparts, as we have shown earlier in another population [9]. This may be an indication that smoking accelerates the degenerative changes of ageing or the degenerative changes induced by smoking itself (e.g. radical formation, lung volume loss, etc.).

Heavy smoking seems to be associated with other lifestyle-related risk factors. In this study, smokers had an increased weekly alcohol consumption compared with nonsmokers, particularly in patients who smoked > 50 pack-years. Alcohol misuse is associated with a highly increased risk of postoperative complications [10]. In our patients, postoperative mortality was tended to increase in smokers with > 50 pack-years; these smokers also had a higher incidence of postoperative pulmonary complications which agrees with other studies [11–13].

Recently, we showed that pre-operative smoking guidelines are highly effective in reducing the incidence of postoperative complications [14]. Pre-operative smoking intervention may be an effective tool for reducing complications in patients with a high risk of postoperative intensive care admission, such as those undergoing colorectal and chest surgery.

In conclusion, patients who smoked heavily have an increased risk of postoperative pulmonary complications and intensive care admission. There seems to be a dose–response relationship between the number of pack-years smoked and postoperative intensive care admission. More research in key areas of the effects of pre-operative smoking intervention on smoking habits and postoperative complications and mortality is required.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Methods
  4. Study population
  5. Study procedure
  6. Statistical analysis
  7. Results
  8. Discussion
  9. Acknowledgements
  10. References

The study was supported by grants from the Danish Medical Research Council, the Danish Ministry of Health and Director Danielsen's Foundation. The authors wish to thank Mrs Jane Cracknell for kindly revising the manuscript.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Study population
  5. Study procedure
  6. Statistical analysis
  7. Results
  8. Discussion
  9. Acknowledgements
  10. References