The Effect of Postoperative Smoking on Femoropopliteal Bypass Grafts

The Effect of Postoperative Smoking on Femoropopliteal Bypass Grafts F. Michael Ameli, MB, ChB, M. Stein, BA, L. Aro, RN, J.L. Provan, MS, FRCS, R. Prosser, MA, Toronto, Ontario, Canada

Effect of smoking habits on limb loss rates and cumulative patency rates of 136 arterial reconstructions performed for lower limb ischemia were analyzed in a five year follow-up retrospective study. Of 121 patients, 103 (85%) smoked before the operation and 43 of the smokers (42%) discontinued smoking postoperatively. Patients who continued to smoke more than 15 cigarettes per day (34 patients) increased the probability of losing their limb approximately five times at two years and three times at five years postoperatively, compared with nonsmokers and smokers of up to 15 cigarettes per day (87 patients) (p = 0.013). Cumulative patency rates of nonsmokers and smokers of up to five cigarettes per day (Group A, 66 patients) were not significantly influenced (p = 0.518) by preoperative symptoms (claudication versus limb salvage). However, for smokers of more than five cigarettes per day (Group B, 55 patients), at five years ciaudicants had a cumuiatlve patency rate of 62.9Yo compared to 38.3% for limb salvage patients (p = 0.015). In group A at five years, autologous saphenous vein grafts had a cumulative patency rate of 74.2%, compared to 35% for prosthetic grafts (p = 0.013). In group 6 the CPR differences between autologous saphenous vein and prosthetic grafts were not significantly different (p = 0.394). Multiple interactions between smoking and variables like age, preoperative symptoms, and graft material demonstrate the complexity of the effects of smoking on cumulative patency rate and the need for sub-grouping and removal of confounding factors. In vlew of the adverse affects of continued smoking on postrevascularization prognosis, patients should be strongly advised to discontinue smoking. KEY WORDS: Smoking; limb loss rate; cumulative patency rate; femoropopliteal bypass

Cigarette smoking is a major risk factor for atherosclerotic peripheral vascular disease [ 1-31. The prevalence of smoking in patients with peripheral vascular disease is higher than in the general population and ranges between 80-90% [4-61. Giving up cigarette smoking was shown to improve the resting ankle systolic pressure and exercise tolerance [7] and improve prognosis [8] of claudicants.

A few atherogenic mechanisms were proposed for smoking. Smokers tend to have higher hemoglobin levels, probably induced by carboxyhemoglobinemia, and as a result, increased blood viscosity, which in turn impairs small vessel flow [6]. Carboxyhemoglobinemia was associated with hypoxia in blood vessel walls that could induce cellular injury and atherosclerotic changes [9]. Nicotine was shown to reduce vascular production to prostacyclin (PG,2) which is a vasoFrom the Division of Vascular Surgery, The Wellesley dilator and an antiaggregating factor [lo]. Smokers Hospital, The University of Toronto, Toronto, Canada. were also shown to have lower levels of serum HDL2 Reprint requests: F. Michael Ameli, MB, 160 Wellesley cholesterol and HDL subfraction. HDL and HDL2 Street East, E.K. Jones Building, Suite 313, Toronto, are associated with reduced risk of coronary artery Ontario, M4Y IJ3, Canada. disease [ 111. 20

VOLUME

3

NO 1 - 1989

SMOKING AFTER FEMOROPOPLITEAL BYPASS

21

This retrospective study was undertaken to examine the effects of smoking on primary and postthrombectomy cumulative patency rates (CPRs) and limb loss rates (LLRs). The effects of interaction between smoking and variables like age, graft material, and operative indications on CPRs were analyzed.

based on pulse examination, Doppler studies and, if necessary, arteriography. For analysis purposes, first postoperative occlusion was considered as graft failure. If occluded, grafts were thrombectomized. We followed up postthrombectomy patency intervals and recorded them separately. Intervals until amputation were recorded regardless of patency status along the way.

MATERIAL AND METHODS

In this study special attention was paid to smoking habits before and after operation. At the first interview with each patient, preoperative smoking was carefully verified and recorded as the numbers of pack-years smoked. Postoperatively, smoking was verified in personal interviews and not through written communication. It was recorded for each patient as the number of cigarettes smoked per day.

From January 1980 to December 1985, 136 limbs (121 patients) were operated on for claudication or severe ischemia. The mean age of our patients was 65 f 10 (SD) years with a range of 34-89 years. The sex distribution was 70 (58%) males and 51 (42Oro) females. Patient characteristics, risk factors, indications for operation, and type of graft material used are listed in Table I.

Five cigarettes per day were found to be the best discriminatory quantity which produced the largest difference in CPRs between the two groups and 15 cigarettes per day were found to be the best discriminatory quantity for LLRs.

Preoperative examinations included pulse palpation by the operating surgeon, Doppler studies, transcutaneous oxygen tensions (PtcO,) and arteriography. When available and of good quality, autologous saphenous vein (ASV) grafts were used preferentially, otherwise polytetrafluoroethylene (PTFE) or umbilical vein (prosthetic grafts) were used instead. Intraoperatively, the immediate success of each graft was confirmed via Doppler flow examinations or, if necessary, by arteriography. Postoperatively, each patient was seen regularly. In the first year of follow-up patients were seen three or four times and, subsequently, once a year. In the course of five years, 10 patients were lost to follow-up. Grafts were recorded as patent only if unequivocal clinical evidence proved patency. This evidence was

All the information about patients, operations and follow-up findings was organized into a computer data file and was analyzed using a computer statistical package [12]. All cumulative rates were analyzed by the standard life-table method [13]. The Spearman rank correlation coefficient (r) [14] was used as a measure of association between non-parametric variables. The log-rank test [13] was used for overall comparisons between cumulative rates. The level of statistical significance was chosen to be p = 0.05. The p value for the log-rank test which compared cumulative rates applies to the whole period of 60 months of followUP.

TABLE I.-Risk factors, operations and grafts

RESULTS

Smoking before operation Smokers Nonsmokers Smoking after operation Nonsmokers 1-5 cigarettes per day 6-15 cigarettes per day 16-25 cigarettes per day > 25 cigarettes per day Diabetes Hypertension Indications for operation (N = 136operations) Claudication Limb salvage Type of graft used (N = 136operations) Autologous saphenous vein PTFE Umbilical vein Dacron

Number 99 22 61 5 21 21 13 16 54

52 84 64

6:' 1

(%) (82) (18)

Before operation 103 patients (85%) smoked. Forty-three of the smokers (42%) stopped smoking, making a total of 62 (51.2%) postoperative nonsmok(50) ers. There was a negative correlation between age and ( 4) postoperative smoking (r = -0.33, p < 0.001) and a (17) (17) weak correlation between pre- and postoperative (11) smoking (r = 0.19, p = 0.016). (13) The patients were grouped into nonsmokers and (45) smokers of up to five cigarettes per day (group A) and those who smoked more than five cigarettes per day (group B). Table I1 show CPRs of the two groups. The (38) five year CPR difference was 7.1% but the overall (62) difference was not statistically significant (log-rank, p > 0.1). Group B was further divided into claudicant versus (47) limb salvage cases and compared CPRs. Claudicants had clearly superior CPRs compared to limb salvage ( 11 patients. As can be seen in Figure 1, the difference

22

ANNALS OF VASCULAR SURGERY

SMOKING AFTER FEMOROPOPLITEAL BYPASS

TABLE 11.-Cumulative patency rate and cumulative limb loss (groups C and D) grouped by postoperative smoking habits

25

Cumulative Cumulative limb loss rate patency rate Group A Group B Group C Group D (N = 76) (N = 60) (N = 99) (N = 37) YO YO O h YO -2

Interval

0-1 month 1-12 months 1-2 years 2-3 years 3-4 vears 4-5 ;ears

90.7 75.4 65.5 63.6 60.8 55.7

90.0 66.7 55.2 52.6 48.6 48.6

2.0 2.0 3.3 6.5 6.5 10.9

2.7 2.7 14.9 22.8 28.1 28.1

A

3 Claudicants

Grouo A Nonsmokers and smokers of uo to five cigarettes .Per dav. Group B Smokers of more than five cigarettes perday Group C Nonsmokers and smokers of up to 15 cigarettes per day Group D Smokers of more than 15 cigarettes per day 'Tests for statistical significance: Cumulative patency difference between Groups A and B was not significant (log rank, p > 0.1); Cumulative limb loss difference between Groups C and D was significant (log rank, p = 0.013).

o Limb salvage

N=27 N-48

0

1

12

24

36

48

60

MONTHS

reached a maximum of 40.1010 at two years and declined to 24.6% at five years of follow-up (log-rank, p = 0.015). When group A was divided into claudi- Fig. 2. Cumulative primary patency rates are shown cants versus limb salvage cases and CPRs were com- for claudicants versus limb salvage cases. Patients nonsmokers or smokers of up to five cigarettes pared, a different pattern emerged (Fig. 2). There was were per day. Standard errors and number of patent grafts a 3.4% difference at five years in favor of the limb- are shown at each interval. salvage subgroup, but the overall difference was not statistically significant (log-rank, p = 0.518). Groups A and B were divided into ASV versus pros- ASV subgroup showed a clear superiority over the thetic subgroups to examine the interaction between prosthetic subgroup (Fig. 3) with a five year difference smoking and the type of graft used. In group A the of 39.2% (log-rank, p = 0.013). In group B (Fig. 4), the overall difference between the ASV and prosthetic subgroup was not statistically significant (log-rank, p = 0.394).

62.Plb

38.3%

20

1

Claudicants

N=24

The cases were then divided according to age into older than 64 years (group l), and 64 years or younger (group 2). Group 1 was further subdivided into nonsmokers and smokers of up to five cigarettes per day (group lB), and CPRs were compared (Fig. 5). Group 1A showed higher CPRs throughout the five year follow-up and the overall difference was statistically significant (log-rank, p = 0.009). When group 2 was divided similarly into groups 2A and 2B and CPRs were compared, the differences were not as marked during four years of follow-up. Only at five years the difference became 30.5% (Fig. 6), however, the overall difference was not statistically significant (logrank, p = 0.268).

In this study 29 cases were thrombectomized. The thrombectomized groups were then divided into groups A and B according to smoking habits (see 0 4 1 / , 12 24 36 48 60 above), and their postthrombectomy CPRs compared. At three years of follow-up group A had a CPR of MONTHS 75% and group B 46.5%. However, because of the Fig. 1 , Cumulative primary patenc rates are shown small number of cases, the difference was not statistifor claudicants versus limb sakage cases. All cally significant (log-rank, p > 0.1). patients smoked more than five cigarettes per day. A comparison was made of LLRs between groups A Standard errors and number of patent grafts are shown at each interval. and B but there was no significant difference. When o

Limb salvage

N-35

VOLUME 3 NO 1 - 1989

23

SMOKING AFTER FEMOROPOPLITEAL BYPASS

35

68.6%

0

ASV

N-36

o

Prosthetic

N-40 I

3

24

36

48

> 5 cigs per day

N-27

60

MONTHS

+

Fig. 3. Cumulative primary patency rates are shown for ASV versus prosthetic grafts. Patients were nonsmokers or smokers of up to five cigarettes per day. Standard errors and number of patent grafts are shown at each interval.

' 1

'

1

12

24

36

,

48

1

i

60

MONTHS

Fig. 5. Cumulative primary patency rates are shown for nonsmokers and smokers of up to five cigarettes per day versus smokers of more than five cigarettes per day. All patients were older than 64 years. Standard errors and number of patent grafts are shown at each interval.

the smoking level was raised from five to 15 cigarettes per day a significant difference was found. Smokers of more than 15 per day lost significantly more limbs compared to smokers of 15 cigarettes or less and nonsmokers (Table 11). There was a difference of 21.6% at four years and 17.2% at five years of follow-up (log-rank, p = 0.013).

im

80

60.3%

Preoperative smoking recorded in pack-years did not significantly affect CPRs or LLRs.

39.1%

DISCUSSION

Y

1 -I4

5

40

J 0 o

0

Prosthetic

N-32

1

12

24

36

48

60

MONTHS

Fig. 4. Cumulative primary patency rates are shown for ASV versus prosthetic grafts. All patients smoked more than five cigarettes per day. Standard errors and number of patent grafts are shown at each interval.

Some studies reported the detrimental effects of postoperative smoking on patency of aortofemoral Dacron grafts [4,15,16], femoropopliteal vein grafts [4,17], seeded femoropopliteal Dacron grafts [181, and aortoiliac Dacron grafts [16,17]. Our results are consistent with the above findings in that smoking more than five cigarettes per day postoperatively, was associated with lower primary- and postthrombectomy CPRs (Table 11). However, these trends fall short of statistical significance. When the smoking quantity was raised to 15 cigarettes per day a statistically significant association between postoperative smoking and LLR was found (Table 11). Patients who continued to

SMOKING AFTER FEMOROPOPLITEAL BYPASS

24

ANNALS OF VASCULAR SURGERY

The effects of smoking on CPRs were quite different for patients older than 64 years (group 1) comlWV

0

o

up to 5 cigs per day

N=24

> 5 cigs per day

N=33

Fig. 6. Cumulative primary patency rates are shown for nonsmokers and smokers of up to five cigarettes per day versus smokers of more than five cigarettes per day. All patients were 64 years of age or younger. Standard errors and number of patent grafts are shown at each interval.

smoke more than 15 cigarettes per day increased the probability of losing their leg approximately five times at two years and three times at five years, postsurgery, compared with nonsmokers and smokers of up to 15 cigarettes per day. Our results showed that CPRs were affected by multiple interactions between postoperative smoking and variables like preoperative symptoms, graft material, and age. It appears that for nonsmokers and smokers of up to five cigarettes per day (group A) CPRs were not significantly influenced by preoperative symptoms (Fig.2). For smokers of more than five cigarettes per day (group B) there was a significant difference in CPRs between claudicants and limb salvage patients (Fig.1). Limb salvage patients who smoked more than five cigarettes per day had the lowest five year CPR (38.3%) and can be considered as a definite risk group for graft failure. There was an interesting interaction between smoking and the type of graft used. In group A, ASV grafts had significantly higher CPRs compared to prosthetic grafts (Fig.3). For group B, the CPR difference between ASV and prosthetic grafts were not significantly different (Fig.4). Smoking affected mainly ASV grafts by lowering the five year CPR from 74.2% (group A) to 60.3% (group B).

ables significant confounding was removed and, in spite of the smaller samples, associations became

association between postoperative smoking of more than five cigarettes per day and lower CPRs (p = 0.045). Severely ischemic patients who smoked postoperatively more than 15 cigarettes per day had a limb salvage rate of 58.5% compared to 89% for nonsmokers and smokers of up to 15 cigarettes per day (p = 0.009). There is strong evidence that smoking is one of the major risk factors in atherosclerosis of lower limb arteries [5,20,22]. Smoking was also associated with increased atherosclerosis found at autopsy [22]. A recent arteriographic study [23] found an association between smoking and increased arteriosclerotic changes in the common femoral arteries. There is also evidence that a process similar to atherosclerosis leads to occlusion of ASV grafts for peripheral vascular disease [24,25]. Smoking as a major risk factor for atherosclerosis is likely to have similar long-term effects on vein and possibly other grafts, as it has on coronary or femoral arteries. The reliability of patients' reports concerning their smoking habits has been questioned [ 5 ] . Also, by relying on verbal communications with our patients, an incorrect assumption was made that one cigarette was equivalent to another in terms of smoke inhalation. One study [26] assessed smoking habits by both verbal communications and measurements of serum thiocyanate. It was found that statements of complete stopping were more reliable than statements of reduction in smoking. In our study the follow-up interviews were held in a friendly and non-judgmental atmosphere. It would be unrealistic to claim that all our patients told the truth, however, we believe that the majority did. The effects of smoking on CPRs were analyzed by comparing patients who smoked more than five cigarettes per day with nonsmokers and smokers of up to five cigarettes

VOLUME 3 NO 1 - 1989

SMOKING AFTER FEMOROPOPLITEAL BYPASS

per day. Patients who were most likely to underestimate their smoking were those who claimed to smoke between one and five cigarettes per day. However, there were only five of these patients (Table I), which means that most of our comparison was between smokers and nonsmokers. In view of the deleterious effects of smoking on the prognosis of bypass grafts, the question is how aggressive should clinicians be in trying to modify smoking behaviors and how likely is an aggressive attitude to succeed. Our results showed that 42% of preoperative smokers no longer smoked postoperatively. Our patients were advised to stop smoking and told the potential consequences of not doing so without any further pressure. This attitude proved beneficial.

CONCLUSION

This study has shown that postoperative smoking affects CPRs through a complex interaction with variables like age, preoperative symptoms, and graft material. Smoking of more than 15 cigarettes per day significantly increased the probability of postoperative limb loss. In view of the adverse effects of continued smoking on postrevascularization prognosis, patients should be strongly advised to discontinue smoking.

REFERENCES 1. KANNEL WB, SHURTLEFF D. The Framingham study: cigarettes and the development of intermittent claudication. Geriatrics 1973;28:61-68. 2. JACOBSEN UK, DICE-PEDERSEN H, GYNTELBERG F. “Risk factors” and manifestations of arteriosclerosis in patients with intermittent claudication compared to normal persons. (Abstract). Danish Med Bull 1984;31:145-148. 3. HUGHSON WG, MANN JI, GARROD A. Intermittent claudication: prevalence and risk factors. Br Med J 1978; 1:1379-1381. 4. MYERS KA,KING RB, SCOTT DF, JOHNSON N, MORRIS PJ. The effect of smoking on the late patency of arterial reconstructions in the legs. Br JSurg 1978;65:267-271. 5 . THOMAS M. Smoking and vascular surgery. Br J Surg 1981; 68:601-604. 6. CASTELDEN WM, FAULKNER K, HOUSIAK AK, WATT A. Hemoglobin, smoking and peripheral vascular disease. J Royal Soc Med 1981; 74.586-590.

25

7. QUICK CRG, COTTON LT. The measured effect of stopping smoking on intermittent claudication. Br J Surg 1982; 69(S):S24-S26. 8. HUGHSON WG, MANN JI, TIBBS DJ, WOODS HF, WALTON I. Intermittent claudication: factors determining outcome. Br Med J 1978;1:1377-1379. 9. SOJKA SG, PROVAN JL. Cigarette smoking and peripheral vascular disease: is carbon monoxide the real culprit? Canad Med Assoc J 1981;125:lO-11. 10. NADLER JI, VELASCO JS, HORTON B. Cigarette smoking inhibits prostacyclin formation. Lancet 1983;1:1248-1250. 1. SHENNAN NM, SEED M, W Y ” V. Variation in serum lipid and lipoprotein levels associated with changes in smoking behavior in non-obese Caucasian males. Atherosclerosis 1985; 58:15-25. 2. SAS Institute Inc. SAS(R) User’s Guide: Statistics, Version 5 Edition. Cary NC: SAS Institute Inc 1985529-557. 3. ELANDT-JOHNSON R, JOHNSON N. Survival models and data analysis. New York: John Wiley and Sons 1980; 83-122, 258-259. 14. HOLLANDER M, WOLFE DA. Non-parametric statistical methods. New York: John Wiley and Sons 1973; 191-192. 15. WRAY R, DePALMA RG, HUBAY CH. Late occlusion of aortofemoral bypass grafts: influence of cigarette smoking. Surgery 1971 ; 70969-973. 16. ROBICSEK F, DAUGHERTY HK, MULLEN DC, MASTERS TN, NARBAY D, SANGER PW. The effect of continued cigarette smoking on the patency of synthetic vascular grafts in Leriche syndrome. J Thorac Cardiovasc Surg 1975; 70:107- 1 12. 17. GREENHALGH RM, LAING SP, COLE PV, TAYLOR BW. Smoking and arterial reconstruction. Br J Surg 1981; 68:605-607. 18. HERRING M, GARDNER A, GLOVER J. Seeding human arterial prosthesis with mechanically derived endothelium. The detrimental effect of smoking. J Vasc Surg 1984;1:279-289. 19. AMELI FM, STEIN M, PROSSER RJ, PROVAN JL. The outcome of femoropopliteal bypass operations for patients with severe ischemia and the effect of cigarette smoking on limb salvage and patency rates. J. Cardiovasc Surg 1989;(in press). 20. READ RC. Systemic effects of smoking. Am J Surg 1985; 148:706-711. 21. Editorial: Smoking and vascular disease. Br Med J 1972;2; 3-4. 22. STRONG JP, RICHARDS ML. Cigarette smoking and atherosclerosis in autopsied men. Atherosclerosis 1976; 23:45 1-476. 23. HYVARINEN S. Arteriographic findings of claudication patients. Ann Clin Res 1984;16:5-45. 24. SZILAGYI DE, HAGEMAN JH, SMITH RF, DALL’OLMO CA. Biologic fate of autogenous vein implants as arterial substitutes -clinical, angiographic and histopathologic observations in femoropopliteal operations for atherosclerosis. Ann Surg 1973; 178:232-244. 25. WALTON KW, SLANEY G, ASHTON F. Atherosclerosis in vascular grafts for peripheral vascular disease. Atherosclerosis 1985;54:49-64. 26. KIRK CJC, LUND VJ, WOOLCOCK NJ3, GREENHALGH RM. The effect of advice to stop smoking on arterial disease patients, assessed by serum thiocyanate levels. J Cardiovasc Surg 1980;21.568-569.

...