Prolonged Injection Time and Light Smoking Decrease the Incidence of Fentanyl-Induced Cough

Lin, Jui-An MD; Yeh, Chun-Chang MD; Lee, Meei-Shyuan DPH; Wu, Ching-Tang MD; Lin, Shinn-Long MD; Wong, Chih-Shung MD, PhD

Section Editor(s): Bovill, James G.

Anesthesia & Analgesia:
doi: 10.1213/01.ANE.0000159161.31276.DB
Anesthetic Pharmacology: Research Report

We designed this study to evaluate the effect of injection time and smoking on fentanyl-induced cough. Four-hundred-fifty ASA class I–II patients, aged 18–80 yr and weighing 40–90 kg, scheduled for elective surgery were included. All patients received fentanyl (100 μg for patients weighing 40–69 kg and 150 μg for patients weighing 70–90 kg for clinical convenience) via the proximal port of a peripheral IV line on the forearm. Patients were randomly assigned to 3 groups of 150 patients each. Patients in Group I received fentanyl injection over 2 s, whereas for patients in Groups II and III the fentanyl was injected at a constant rate over 15 s and 30 s, respectively. We recorded the number of coughs of each patient during and 30 s after fentanyl injection. The incidence of cough was 18% in group I, 8% in Group II, and 1.3% in Group III, significantly less (P < 0.05) with a longer injection time. Current smokers had a less frequent incidence of cough than nonsmokers; however, this effect was only significant in light smokers (<10 cigarettes per day or <10 smoking years or <10 pack-years). In conclusion, a longer injection time reduces the incidence of fentanyl-induced cough, and light smoking may be a protective factor against fentanyl-induced cough.

In Brief

IMPLICATIONS: Slow injection, particularly at 30 s, minimizes the incidence of fentanyl-induced cough, and light smoking provides a protective effect against cough.

Author Information

Department of Anesthesiology, Tri-Service General Hospital and National Defense Medical Center, National Defense University, Taipei, Taiwan, Republic of China

Accepted for publication January 26, 2005.

Address correspondence and reprint requests to Chih-Shung Wong, MD, PhD, Department of Anesthesiology, Tri-Service General Hospital and National Defense Medical Center, National Defense University. #325, Section 2, Chenggung Road, Neihu 114, Taipei, Taiwan, Republic of China. Address e-mail to

Article Outline

Because it induces intense analgesia with a rapid onset and short duration, decreased cardiovascular depression, and no histamine release (1,2), fentanyl is widely used as an analgesic during induction of general anesthesia. Reflex coughing after IV bolus of fentanyl was reported in various controlled studies (3–7) but has not been considered a serious anesthetic complication. Coughing during anesthesia induction is undesirable and is associated with increased intracranial (ICP), intraocular, and intraabdominal pressures. Prevention of fentanyl-induced cough in such conditions is of clinical importance. Furthermore, Tweed and Dakin (8) reported a case of explosive, spasmodic coughing after peripheral IV injection of fentanyl that was severe enough to produce periorbital petechiae and was only relieved after induction of general anesthesia.

Various attempts have been made to reduce the incidence of fentanyl-induced cough during anesthesia induction (3–5). Our preliminary observations suggested that the incidence could be reduced by prolonging the injection time and we therefore wished to examine this in detail. In addition, smoking leads to respiratory symptoms, such as cough, increased mucus secretion, and airway inflammation. In our routine anesthesia management, the incidence of fentanyl-induced cough does not seem to be more frequent in current smokers, contradicting expectations regarding the impact of smoking on cough. Studies on fentanyl-induced cough have always excluded smokers from analysis and the effects of smoking have not been evaluated. This study was therefore performed to examine the effect of injection time and smoking history on the occurrence of fentanyl-induced cough.

Back to Top | Article Outline


After approval by the Institutional Research and Ethics Committees, informed consent was obtained from all patients included in this randomized, prospective, controlled study. Four-hundred-fifty ASA physical status I–II patients of either sex, aged 18–80 yr and weighing 40–90 kg, scheduled for elective surgery under general anesthesia were randomly assigned to 3 groups of 150 patients each. Before induction of anesthesia, an IV cannula on the dorsum of the patient’s hand was established and connected to a T-connector for drug injection to minimize the dilution effect during fentanyl administration. All patients were given fentanyl (100 μg for those weighing 40–69 kg and 150 μg for those weighing 70–90 kg) via the T-connector with the IV fluid running at a fast rate. Group I patients received the fentanyl injection over 2 s, whereas for patients in Groups II and III, the fentanyl injection was given over a period of 15 or 30 s, respectively. A stopwatch was used to monitor the injection time. We recorded the number of coughs of each patient during and 30 s after fentanyl injection, and the number of coughs was counted by the same anesthesiologist who injected fentanyl. Anesthesia induction was commenced immediately after cough cessation or 30 s after the end of injection. Exclusion criteria included a history of asthma, chronic cough, or upper respiratory tract infection during the previous 2 wk or recent treatment with angiotensin-converting enzyme inhibitors, bronchodilators, or steroids.

Monitoring consisted of an electrocardiogram, noninvasive arterial blood pressure monitoring, pulse oximetry, and capnography. After fentanyl injection, the incidence and severity of cough were recorded by another anesthesiologist who was blinded to the method of fentanyl injection. In a pilot study of 50 patients using a 3-min observation period, we found that all coughs occurred within the 30 s after fentanyl injection, so we chose 30 s as the observation period for the current study. Based on the number of coughs observed, cough severity was graded as mild (1–2), moderate (3–5), or severe (>5). Arterial blood pressure was recorded before fentanyl injection (systolic blood pressure [SBP]1 and diastolic blood pressure [DBP]1) and 30 s after fentanyl injection for patients who did not cough or immediately after cough stopped (SBP2 and DBP2). General anesthesia was induced 30 s after fentanyl injection.

We assessed the smoking status of the patient in several ways. We defined smokers as someone who smoked and did not have chronic airway disorder, such as asthma, chronic obstructive pulmonary disease, or airway obstruction. A current smoker was defined as someone who was still smoking and former smokers as those who had stopped smoking for more than 6 mo before being included in the study; nonsmokers consisted of patients who had never smoked and former smokers. To consider the intensity of smoking, we classified intensity in 3 levels; none, ≤10, and >10 cigarettes per day, pack-years, or duration in years (pack-year indicates the cumulative dose, e.g., 1 pack/day for 10 yr = 2 packs/day for 5 yr = 10 pack-years). Patients smoking more than 10 cigarettes per day, or >10 smoking years, or 10 pack-years were defined as heavy smokers.

All data are reported as mean ± sd or percentages. The characteristics of the subjects in the three groups were evaluated by one-way analysis of variance to compare the means and by χ2 tests to compare the incidence of events. Tukey tests were applied to compare the difference among groups if analysis of variance was significant. The log-rank test and Cox proportional hazard model were used to evaluate differences in time to cough. Logistic regression analyses were used to evaluate the relationship between fentanyl-induced cough and injection time and/or smoking status. Estimates of the odds ratios (OR) and associated 95% confidence intervals (CI) were obtained from these models. All statistical analyses were two tailed. Statistical significance was accepted at the 5% probability level.

Back to Top | Article Outline


The patient characteristics were comparable in the 3 groups (Table 1). There were more males than females in each group, but there was no difference in gender distribution among the 3 groups. Patients in Group I had significantly higher SBP2 and DBP2 values, but not SBP1 and DBP1 values, than the other 2 treatment groups (P < 0.05) (SBP2: 127.9 mm Hg versus 120.9 and 121.0 mm Hg; DBP2: 76.7 mm Hg versus 69.9 and 70.9 mm Hg), but there was no clinical significance. To further clarify the increasing of SBP2/DBP2 between the patients in Group I who coughed or did not cough, we found a significant increase of the SBP2/DBP2 in patients who coughed compared with those who did not cough, the SBP2/DBP2 values were 147.7 ± 13.0/89.4 ± 16.8 versus 123.5 ± 29.3/73.8 ± 10.4 mm Hg (P < 0.001).

The incidence of fentanyl-induced cough was 18% in Group I, 8.0% in Group II, and 1.3% in Group III (Table 2). Patients in Groups II and III had a significantly less frequent incidence of cough than those in group I. Moreover, the longer the injection time, the less frequent the incidence of cough. In patients with cough, moderate to severe cough was seen in 40.7% (11 of 27) of patients in Group I and in 50.0% (6/12) in Group II but in none in Group III. There was no significant difference in the time of onset of coughing among the groups (data not shown).

No significant difference was seen among the groups in the distribution of smoking status (never smoker, former smoker, current smoker) (Table 1) or intensity of smoking (cigarette/day; smoking duration; pack-year). Current smokers showed a less frequent fentanyl-induced cough than nonsmokers in all 3 groups (P < 0.05) (Fig. 1).

We used four separate measures of smoking intensity to predict cough by using logistic regression (Table 3). Current smokers who smoked fewer than 10 cigarettes per day had less fentanyl-induced cough than nonsmokers (OR = 0.12; 95% CI, 0.02–0.86; P = 0.04), but there was no difference between heavy smokers and nonsmokers (OR = 0.34; 95% CI, 0.10–1.12; P = 0.08). Similarly, current smokers who had smoked for <10 yr also showed less frequent cough than nonsmokers (OR = 0.08; 95% CI, 0.01–0.62; P = 0.02), and there was no difference between current smokers who had smoked for more than 10 yr and nonsmokers (OR = 0.52; 95% CI, 0.15–1.77; P = 0.30). Moreover, current smokers with a dosage of <10 pack-years had less frequent cough than nonsmokers (OR = 0.15; 95% CI, 0.04–0.64; P = 0.01), but this difference was not seen in current smokers with a dosage more than 10 pack-years compared with nonsmokers (OR = 0.45; 95% CI, 0.11–1.97; P = 0.29). Thus, regardless of the method used to estimate intensity of smoking, the data consistently showed that cough incidence was significantly decreased in the light smoker group (dosage ≤10) compared with the nonsmoker group.

To confirm the effect of injection time and smoking, we also included both variables in the model. After controlling the two variables injection time and smoking for each other (Table 4), both were found to have independent effects on fentanyl-induced cough.

Back to Top | Article Outline


Fentanyl, administered via a peripheral IV cannula, provoked cough in 18% of patients when injected within 2 seconds, but the incidence decreased significantly as the injection time was increased. In addition, current smokers showed a less frequent incidence of cough in all three groups compared with nonsmokers, whereas heavy smokers did not.

Previous clinical studies (3–7) on cough induced by fentanyl injection via a peripheral line without premedicants found a variable incidence (Table 5). This discrepancy can be explained by the differences in injection dose and time. Although a larger dose and shorter injection time were used in the study by Bohrer et al. (7), cough incidence was less than in the other reports. Lin et al. (4) suggested that age could be a confounding factor for fentanyl-induced cough, as the average age in Bohrer’s group (7) was more than 60 years, which could be explained by possible heightened irritant receptor activity in the younger population. However, no effect of age on cough incidence was seen in our study by analysis of all patients together or nonsmokers alone (data not shown). In our study, the incidence of fentanyl-induced cough was minimized simply by slowing the rate of injection without any drug pretreatment. From the pharmacokinetic point of view, the duration of drug injection may affect the peak plasma concentration, with a longer injection time resulting in a smaller peak concentration. The threshold for fentanyl-induced cough may be reached more easily at a larger peak plasma concentration; therefore, the longer the injection time, the less frequent fentanyl-induced cough. We also found that slowing the rate of injection reduced not only the incidence but also the severity of cough (Table 2). Moreover, patients without cough after fentanyl injection maintained a stable hemodynamic status.

A specific receptor, known as the rapidly adapting receptor, is thought to be the cough receptor. However, in an in vitro study, Fox (9) found that rapidly adapting receptors are insensitive to tussigenic stimuli. It has been suggested that pulmonary C-fiber receptor activation can directly cause cough (10). C-fibers are readily activated by chemical stimuli (11,12) and may release tachykinins, which cause secondary mucosal responses and excite the rapidly adapting receptors. In the present study, we found that current smokers had less cough than nonsmokers and that this was independent of gender or injection time. However, the benefit of smoking was only observed in current light smokers but not in heavy smokers. We suggest that nicotine and airway hyperresponsiveness may play important roles in the impact of smoking on fentanyl-induced cough. Chronic tobacco exposure augments the substance P-evoked increase in activity of the rapidly adapting receptors and the irritant receptor and thus induces airway hyperresponsiveness (13). Smoking cessation clearly improves airway hyperresponsiveness (14), which is closely related to cough (15,16). In contrast, cough sensitivity is decreased in smokers, supporting the hypothesis that nicotine inhibits or blocks C-fiber activity in the sensory nervous system of the lower respiratory tract (17). Taken together, nicotine can excite rapidly adapting receptors (18,19) and inhibit C-fiber activity. The protective effect of nicotine on fentanyl-induced cough, in the present study, can be explained by nicotine inducing less excitatory effect on rapidly adapting receptors and more inhibitory effect on C-fibers in current light smokers. Besides, heavy smokers have a more sensitive airway than light smokers and the inhibitory effect of nicotine on coughing in current smokers might be abolished in heavy smokers with more hypersensitive airways. These results suggest that fentanyl-induced cough acts mainly via the C-fiber mechanism, which is consistent with the suggestion of Bohrer et al. (7).

In conclusion, fentanyl injection (≈2 μg/kg) induced cough in 18% of patients when the injection time was <2 seconds, and the incidence of evoked cough decreased with injection time to 1.3% for an injection time of 30 seconds. Fentanyl-induced cough is a problem in some situations, such as ruptured eyeball or increased ICP; slow injection of fentanyl can almost obviate the problem and make the induction smoother. Also, smoking may have a protective effect against fentanyl-induced cough in light smokers but not in heavy smokers.

Back to Top | Article Outline


1. Bovill JG, Sebel PS, Stanley TH. Opioid analgesics in anesthesia: With special reference to their use in cardiovascular anesthesia. Anesthesiology 1984;61:731–55.
2. Grell FL, Koons RA, Denson JS. Fentanyl in anesthesia: A report of 500 cases. Anesth Analg 1970;49:523–32.
3. Agarwal A, Azim A, Ambesh S, et al. Salbutamol, beclomethasone or sodium chromoglycate suppress coughing induced by IV fentanyl. Can J Anaesth 2003;50:297–300.
4. Lin CS, Sun WZ, Chan WH, et al. Intravenous lidocaine and ephedrine, but not propofol, suppress fentanyl-induced cough. Can J Anaesth 2004;51:654–9.
5. Lui PW, Hsing CH, Chu YC. Terbutaline inhalation suppresses fentanyl-induced coughing. Can J Anaesth 1996;43:1216–9.
6. Phua WT, Teh BT, Jong W, et al. Tussive effect of a fentanyl bolus. Can J Anaesth 1991;38:330–4.
7. Bohrer H, Fleischer F, Werning P. Tussive effect of a fentanyl bolus administered through a central venous catheter. Anaesthesia 1990;45:18–21.
8. Tweed WA, Dakin D. Explosive coughing after bolus fentanyl injection. Anesth Analg 2001;92:1442–3.
9. Fox AJ. Modulation of cough and airway sensory fibres. Pulm Pharmacol 1996;9:335–42.
10. Karlsson JA, Fuller RW. Pharmacological regulation of the cough reflex: From experimental models to antitussive effects in man. Pulm Pharmacol Ther 1999;12:215–28.
11. Sant’Ambrogio FB, Sant’Ambrogio G. Circulatory accessibility of nervous receptors localized in the tracheobronchial tree. Respir Physiol 1982;49:49–73.
12. Paintal AS. Mechanism of stimulation of type J pulmonary receptors. J Physiol 1969;203:511–32.
13. Bonham AC, Kott KS, Joad JP. Sidestream smoke exposure enhances rapidly adapting receptor responses to substance P in young guinea pigs. J Appl Physiol 1996;81:1715–22.
14. Willemse BW, Postma DS, Timens W, ten Hacken NH. The impact of smoking cessation on respiratory symptoms, lung function, airway hyperresponsiveness and inflammation. Eur Respir J 2004;23:464–76.
15. Hsiue TR, Hsieh AL, Chang HY, et al. Bronchoprovocation test by forced oscillation technique: Airway hyperresponsiveness in chronic cough and psychogenic dyspnea subjects. J Formos Med Assoc 1993;92:231–6.
16. Chang AB, Phelan PD, Sawyer SM, Robertson CF. Airway hyperresponsiveness and cough-receptor sensitivity in children with recurrent cough. Am J Respir Crit Care Med 1997;155:1935–9.
17. Millqvist E, Bende M. Capsaicin cough sensitivity is decreased in smokers. Respir Med 2001;95:19–21.
18. Zhang Z, Bonham AC. Lung congestion augments the responses of cells in the rapidly adapting receptor pathway to cigarette smoke in rabbit. J Physiol 1995;484:189–200.
19. Ravi K, Kappagoda CT, Bonham AC. Pulmonary congestion enhances responses of lung rapidly adapting receptors to cigarette smoke in rabbit. J Appl Physiol 1994;77:2633–40.
© 2005 International Anesthesia Research Society