The Effect of Smoking Cessation and Subsequent Resumption on Absorption of Inhaled Insulin

Emerging Treatments and Technologies ORIGINAL ARTICLE

1 1 REINHARD H.A. BECKER, MD, PHD ANNKE D. FRICK, PHD cutaneous insulin therapy (6). 2 2 SUE SHA, MD, PHD ROBERT J. FOUNTAINE, PHARMD Furthermore, the availability of inhaled insulin could, in theory, help to overcome patient aversion to insulin therapy in gen- OBJECTIVE — To assess the absorption proﬁle of inhaled insulin in healthy, actively smok- eral (7). ing subjects at baseline, after smoking cessation, and after smoking resumption and compare it Pulmonary permeability determines with nonsmoking subjects. the absorption of inhaled peptides, and smoking has been shown to increase the RESEARCH DESIGN AND METHODS — Insulin pharmacokinetics and glucodynam- permeability of the lungs to diethylenetri- ics were measured in 20 male smoking subjects (10–20 cigarettes/day) and 10 matched non- amine penta-acetic acid radioaerosols (8– smoking subjects after receiving inhaled insulin (1 mg) or the approximate subcutaneous insulin 11), an effect that is independent of the equivalent (3 units) in a randomized cross-over fashion. All smokers then received inhaled insulin 12 h, 3 days, and 7 days into a smoking cessation period. They then resumed smoking for presence of nicotine in the blood (12). 2–3 days before again receiving inhaled insulin 1 h after the last cigarette. Accordingly, smoking has been shown speciﬁcally to increase inhaled insulin ab-

RESULTS — Before smoking cessation, maximum insulin concentration (Cmax) and area un- sorption in healthy volunteers (13–16). der the curve (AUC) for insulin concentration time (AUC-Insulin0–360) with inhaled insulin Therefore, the nicotine-independent ef- were higher, and time to Cmax (tmax) shorter, in smokers than nonsmokers (Cmax 26.8 vs. 9.7 fects of tobacco smoke on the lung may ␮ ␮ Ϫ1 Ϫ1 U/ml; AUC-Insulin0–360 2,583 vs. 1,645 U ml min ; tmax 20 vs. 53 min, respectively; have implications for inhaled insulin all P Ͻ 0.05), whereas with subcutaneous insulin, systemic exposure was unchanged (AUC- therapy. ␮ Ϫ1 Ϫ1 ϭ Insulin0–360 2,324 vs. 2,269 U ml min ; P NS). After smoking cessation, AUC- In smokers, cessation of smoking sig- Insulin decreased with inhaled insulin by up to 50% within 1 week and approached 0–360 nificantly decreases the clearance of ra- nonsmoker levels. Cmax decreased and tmax increased relative to baseline but were still not comparable with nonsmoker values. Smoking resumption completely reversed the effect of dioaerosol particles from the lung and smoking cessation. Glucodynamics corroborated the observed findings in insulin pharmacoki- decreases permeability within 1–3 weeks; netics. however, complete reversal of the increased permeability in smoking relative CONCLUSIONS — Cessation and resumption of smoking greatly altered the pharmacoki- to nonsmoking subjects is not achieved netics of inhaled insulin. As rapid changes in systemic insulin exposure increase hypoglycemia within that period (9,17). Nevertheless, a risk, inhaled insulin should not be used in people with diabetes who choose to continue smok- significant impact on alveolar permeabil- ing. This is consistent with recommendations that people with diabetes refrain from smoking ity can be observed after only 12–24 h altogether. without smoking (17,18). Recent studies with inhaled insulins show that 3–4 Diabetes Care 29:277–282, 2006 weeks of smoking cessation restores inhaled insulin absorption kinetics and glu- nhaled insulin is being developed as an tients with diabetes. Accordingly, the re- codynamic responses toward those of alternative method of insulin adminis- sults of large-scale clinical trials show that healthy nonsmoking volunteers (15,19), I tration. It has a faster onset of action inhaled insulin is effective and well toler- an effect that does not change further after than either regular human insulin or in- ated in patients with type 1 and type 2 3 months of smoking cessation (15). sulin lispro injected subcutaneously but diabetes and that it may prove to be a However, the time course of the onset and retains a duration of action that is longer novel and well-accepted component of offset of the effects of smoking cessation than that of short-acting analogs of hu- diabetes therapy for many patients (2–5). and resumption on inhaled insulin ab- Ͼ man insulin (1). These characteristics In long-term extension trials, 80% of sorption is currently not known. The pur- suggest that inhaled insulin is suitable for insulin-treated patients preferred an in- pose of this study was 1) to investigate the prandial insulin supplementation in pa- haled insulin regimen over standard sub- absorption profile of inhaled insulin ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● (Exubera) in healthy nondiabetic active From 1Aventis Pharma Deutschland, Frankfurt, Germany; and 2Pfizer Global Research and Development, chronic smokers at baseline, after smok- Groton, Connecticut. ing cessation, and after smoking resump- Address correspondence and reprint requests to Reinhard H.A. Becker, Clinical Discovery and Human Pharmacology, Aventis Pharma Deutschland, Industriepark Ho¨chst, Building H831, Room C0441, 65926 tion, and 2) to compare inhaled insulin Frankfurt am Main, Germany. E-mail: reinhard.becker@sanofi-aventis.com. absorption in smokers with that in Received for publication 7 October 2005 and accepted in revised form 11 November 2005. healthy nonsmokers. Abbreviations: AUC, area under the curve. A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion RESEARCH DESIGN AND factors for many substances. METHODS — This study was carried © 2006 by the American Diabetes Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby out according to the principles of the Dec- marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. laration of Helsinki and Good Clinical

DIABETES CARE, VOLUME 29, NUMBER 2, FEBRUARY 2006 277 Smoking and inhaled insulin absorption

Practice and in keeping with local legal (volume 240 ml) of the Exubera Pulmo- tmax was analyzed by nonparametric anal- and regulatory requirements in Germany. nary Delivery System (Nektar Therapeu- ysis. For the comparison of inhaled and It was approved by the local ethics tics, San Carlos, CA); 1 mg Exubera subcutaneous insulin in part 1 of the committee. contains 27.5 units human insulin. Sub- study (i.e., INH1Sm versus SCSm and This single-center, open-label study jects were instructed to inhale steadily INH1NSm versus SCNSm), an ANOVA included two study groups: 1) 20 active and deeply after a normal exhalation and model with subject, treatment, and pe- chronic smokers willing to quit smoking to hold their breath for 5 s afterward. riod effects was used to analyze ln(AUC) for 7 days and then resume again and 2) and ln(Cmax); tmax was analyzed by non- 10 matched nonsmokers. Inclusion crite- Pharmacokinetic and glucodynamic parametric analysis. For the mean ratios ria included a normal oral glucose toler- evaluations of all pairwise group comparisons for ance test, predose fasting blood glucose An intravenous cannula was inserted into AUC and Cmax, 90% CIs were calculated; Ն Ͻ 80 mg/dl, and HbA1c 6.4%. In addi- the forearm to collect blood samples for for tmax, nominal 90% nonparametric CIs tion, the lung function parameters of determination of blood glucose, immuno- for the respective median difference in forced expiratory volume in1sand reactive serum insulin, plasma carboxy- groups were calculated based on the forced vital capacity had to be Ͼ75% of hemoglobin, and serum cotinine. Hodges-Lehmann estimator. Relative bio- predicted values. All subjects provided Samples (sufficient to provide 2 ml se- availability was calculated as the ratio of written informed consent. rum) for glucose, insulin, and C-peptide the dose-adjusted AUCs (between 0 and The study consisted of three parts. In measurement were taken 60, 30, and 15 360 min) for inhaled and subcutaneous the first part, all subjects (smokers and min before insulin dosing and predose, 5, insulin (i.e., AUCINH/DoseINH)/(AUCSC/ nonsmokers) received a single 1-mg dose 10, 15, 20, 30, 45, 60, 75, 90, 120, 180, DoseSC). of inhaled insulin and a single 3-unit dose 240, 300, and 360 min after administra- For glucodynamic analysis, AUC for of subcutaneous regular human insulin, tion. Samples for carboxyhemoglobin (1 change in blood glucose concentration separated by at least 6 days, in a random- ml of EDTA whole blood) and cotinine (2 (AUC-GLU0–360), maximum change in ⌬ ized cross-over fashion. This established ml blood for 1 ml serum) were taken 60 blood glucose ( GLUmax), and time to ⌬ ⌬ the inhaled insulin baseline pharmacoki- min before study medication on all dosing GLUmax (t GLUmax) were calculated to netics/glucodynamics for smokers days. Immunoreactive serum insulin and compare the extent of blood glucose ex- (INH1Sm) and nonsmokers (INH1NSm) serum C-peptide were determined by ra- cursions. Statistical comparisons were and was the only part in which nonsmok- dioimmunoassays (Prof. Hans K.L. identical to those used for pharmacoki- ers participated. All smokers then pro- Hundt, FARMOVS-PAREXEL, Bloemfon- netic data. ceeded to part 2 of the study, which tein, South Africa). A rat insulin radioim- consisted of three sequential treatment munoassay kit from Linco Research (St. RESULTS — Twenty otherwise stages where a single dose of inhaled in- Charles, MO) was used along with a hu- healthy nondiabetic male smokers (mean sulin was administered on day 1 man insulin calibration standard and age 28.4 years [range 21–44], BMI 22.4 2 (INH2Sm, 12 h after smoking cessation), quality control samples prepared by kg/m [20–26]) and 10 matched non- day 3 (INH3Sm, 3 days after smoking ces- FARMOVS-PAREXEL. The C-peptide kit smokers (mean age 28.2 years [19–37], 2 sation), and day 7 (INH4Sm, 7 days after (C-PEP-RIA CT) was supplied by Bio- BMI 24.5 kg/m [19–28]) participated in smoking cessation). During these 7 days, Source Europe (Nivelles, Belgium). Blood this study. Smokers had been smoking subjects were not allowed to smoke until glucose was determined by glucometer 10–20 cigarettes per day (mean 16.4) for the final blood sample was taken on day 7. (GlucoTouch; Lifescan, Milpitas, CA). an average of 11.4 years (range 2–28). In part 3, subjects resumed smoking (20 Smoking status and adherence to smok- cigarettes/day) until receiving a single Statistical methods ing protocols was confirmed by a car- dose of inhaled insulin on either day 9 or For pharmacokinetic analysis, area under boxyhemoglobin level Ͼ1.9% and serum Ͼ 10 (INH5Sm, 2 or 3 days after resumption the curve (AUC) for serum insulin con- cotinine concentrations 100 ng/ml at of smoking). For parts 1 and 3, smokers centration time between 0 and 360 min screening and during the randomized had their last cigarette 1 h before, and did (AUC-Insulin0–360; calculated by the lin- two-way crossover baseline study (Table not resume smoking until 3 h after, the ear trapezoidal rule), maximum concen- 1). The nonsmoking group consisted of Ն administration of study medication. For tration (Cmax), and time to Cmax (tmax) four ex-smokers (no smoking for 8 part 2, smokers had their last cigarette were determined for each subject and in- months) and six subjects who had never 12 h before INH2Sm and then ceased sulin administration. Exogenous insulin smoked. Nonsmoking status was con- smoking until 6 hours after INH4Sm. All concentrations were derived using C- firmed by a carboxyhemoglobin level smokers were hospitalized for the dura- peptide correction. Statistical compari- Ͻ1.2% and serum cotinine concentration of parts 2 and 3 of the study. sons among smokers over time (i.e., tions below the detection limit of 10 Regular human insulin (HOE31 HPR INH1Sm and INH2–5Sm) used the paired ng/ml in nonsmokers. 100; Aventis Pharma Deutschland, Student’s t test on ln(AUC) and ln(Cmax) Frankfurt, Germany) was administered (i.e., natural log-transformed data); un- Randomized two-way crossover by subcutaneous injection into a lifted ab- transformed tmax was analyzed using the baseline (part 1) dominal skinfold by a health care profes- Wilcoxon’s signed-rank test. For the com- Systemic insulin exposure (AUC- sional using a disposable syringe (Micro- parison of smokers versus nonsmokers in Insulin0–360) with inhaled insulin was Fine 0.3 ml 28G 1/2; Becton Dickinson, part 1 (i.e., INH1Sm versus INH1NSm and significantly higher in smokers compared ϭ Franklin Lakes, NJ). Inhaled insulin (1 subcutaneous [SC]Sm versus SCNSm), an with nonsmokers (INHSm/INHNSm mg Exubera) was administered as an aero- ANOVA model with group effects was 1.57) (Table 2, Fig. 1A). Systemic insulin solized cloud from the holding chamber used to analyze ln(AUC) and ln(Cmax); exposure after subcutaneous insulin, on

278 DIABETES CARE, VOLUME 29, NUMBER 2, FEBRUARY 2006 Becker and Associates

ϭ Table 1—Smoking status at screening and prior to study visits, as indicated by serum cotinine INH1Sm 0.90), Cmax was significantly ϭ and plasma carboxyhemoglobin levels lower (INH3Sm/INH1Sm 0.69), and tmax was significantly longer (median 30.0 Serum cotinine level (ng/ml) Carboxyhemoglobin (%) min). After 7 days of smoking cessation (INH4 ), the decrease in total systemic ϭ Sm Nonsmokers (n 10) insulin exposure had become significant Ͻ Ͻ Screening 10 ( 10) 0.6 (0.4–1.1) relative to baseline (INH4 /INH1 ϭ Ͻ Ͻ Sm Sm Cross-over visit 1 10 ( 10) 0.2 (0.2–0.8) 0.73), such that it became comparable Ͻ Ͻ Cross-over visit 2 10 ( 10–15) 0.2 (0.2–1.2) with nonsmokers (INH1 /INH4 ϭ Smokers (n ϭ 20) NSm Sm 0.87). Cmax was reduced further Screening 303 (120–466) 4.8 (2.7–9.2) ϭ (INH4Sm/INH1Sm 0.59) and tmax de- Cross-over visit 1 297 (149–522) 3.8 (1.9–15.2) layed further (median 37.5 min). These Cross-over visit 2 282 (184–489) 3.9 (2.2–6.6) values approached those of nonsmokers, INH2Sm visit 276 (124–494) 2.3 (1.6–3.9) Ͻ Ͻ such that the difference in tmax relative to INH3Sm visit 29 ( 10–71) 0.8 (0.4–2.9) Ͻ Ͻ nonsmokers became nominal, although INH4Sm visit 10 ( 10–13) 0.5 (0.9–1.4) INH5 visit 243 (114–402) 2.7 (1.6–3.8) Cmax was still significantly higher Sm (INH1 /INH4 ϭ 0.61). Resumption Data are median (range). NSm Sm of smoking for 2–3 days (INH5Sm)restored total systemic insulin exposure ϭ the other hand, was not significantly dif- Smoking cessation and resumption (INH5Sm/INH1Sm 1.22) and Cmax ϭ ferent (SC /SC ϭ 0.98). Maximum (parts 2 and 3) (INH5Sm/INH1Sm 1.09) to the smoker Sm NSm baseline, although t remained signifi- insulin concentration (Cmax) after inhaled Twelve hours after smoking ceased max cantly delayed at 30.0 min. insulin was about three times higher in (INH2Sm), total systemic insulin exposure ϭ Adherence to smoking cessation was smokers (INHSm/INHNSm 2.77), with inhaled insulin was significantly in- confirmed by carboxyhemoglobin levels, whereas Cmax after subcutaneous insulin creased compared with baseline ϭ was not significantly different in smokers (INH2Sm/INH1Sm 1.22). Cmax was also which declined rapidly within 12 h to half ϭ ϭ Ͼ (SCSm/SCNSm 1.17). Time to tmax after increased (INH2Sm/INH1Sm 1.22), of the 1.9% baseline level and to non- inhaled insulin was significantly shorter whereas tmax remained unchanged at 20.0 smoking levels within 3 days of smoking in smokers than in nonsmokers (median min (Table 2, Fig. 1B). However, after 3 cessation (Table 1). Serum cotinine con- 20.0 vs. 52.5 min), whereas after subcu- days of smoking cessation (INH3Sm), total centrations remained almost unchanged taneous insulin, tmax was not significantly systemic insulin exposure had started to 12 h after cessation of smoking (in line different (median 90 min in both cases). decline relative to baseline (INH3Sm/ with the approximate 24-h half-life of nic-

Table 2—Pharmacokinetic summary measures (C-peptide corrected) after inhalation of 1 mg insulin or subcutaneous injection of 3 units regular human insulin in nondiabetic nonsmokers and smokers at baseline, after smoking cessation, and after smoking resumption

AUC-Insulin0–360 ␮ Ϫ1 Ϫ1 ␮ ( U ml min )* Cmax ( U/ml)* tmax (min)†

SCNSm 2,324 11.6 90.0 (127.5) SCSm 2,269 13.6 90.0 (97.5)

INH1NSm 1,645 9.7 52.5 (55.0) INH1Sm 2,583 26.8 20.0 (21.8) INH2Sm 3,165 33.3 20.0 (22.5)

INH3Sm 2,321 18.5 30.0 (37.8) INH4Sm 1,887 15.8 37.5 (41.8) INH5Sm 3,156 29.2 30.0 (30.3)

Mean AUC ratio (90% CI) Mean Cmax ratio (90% CI) tmax median difference (90% CI) ϭ ϭ Ϫ Ϫ ϭ SCSm/SCNSm 0.98 (0.82–1.16); P NS 1.17 (0.94–1.45); P NS 15.0 ( 60.0 to 15.0); P NS Ͻ Ͻ Ϫ Ϫ Ϫ Ͻ INH1Sm/INH1NSm 1.57 (1.15–2.14); P 0.05 2.77 (1.99–3.85); P 0.001 30.0 ( 55.0 to 10.0); P 0.01 ϭ ϭ Ϫ Ϫ Ϫ Ͻ INH1NSm/SCNSm 0.71 (0.52–0.97); P NS 0.83 (0.64–1.08); P NS 57.5 ( 135.0 to 15.0); P 0.05 ϭ Ͻ Ϫ Ϫ Ϫ Ͻ INH1Sm/SCSm 1.14 (0.96–1.35); P NS 1.97 (1.60–2.42); P 0.001 70.0 ( 80.0 to 55.0); P 0.001 Ͻ Ͻ Ϫ ϭ INH2Sm/INH1Sm 1.22 (1.04–1.45); P 0.05 1.24 (1.06–1.46); P 0.05 2.5 ( 5.0 to 10.0); P NS ϭ Ͻ Ͻ INH3Sm/INH1Sm 0.90 (0.74–1.08); P NS 0.69 (0.58–0.82); P 0.01 10.0 (10.0–25.0); P 0.001 Ͻ Ͻ Ͻ INH4Sm/INH1Sm 0.73 (0.60–0.88); P 0.05 0.59 (0.50–0.69); P 0.001 17.5 (10.0–30.0); P 0.01 ϭ ϭ Ͻ INH5Sm/INH1Sm 1.22 (1.03–1.45); P NS 1.09 (0.91–1.31); P NS 10.0 (0.0–15.0); P 0.01 ϭ Ͻ Ϫ ϭ INH1NSm/INH4Sm 0.87 (0.62–1.22); P NS 0.61 (0.47–0.80); P 0.01 15.0 ( 10.0–30.0); P NS ϭ AUC-Insulin0–360 area under the serum insulin concentration versus time curve between 0 and 360 min. *Geometric mean values; †Median (arithmetic mean) values.

DIABETES CARE, VOLUME 29, NUMBER 2, FEBRUARY 2006 279 Smoking and inhaled insulin absorption

Figure 1—Serum insulin concentration time proﬁles after inhalation of 1 mg insulin (INH) or subcutaneous injection of 3 units regular human insulin (SC) in nondiabetic smokers (Sm; n ϭ 20) and nonsmokers (NSm; n ϭ 10). A: Baseline cross-over study. B: After smoking cessation Ϯ (INH2Sm-INH4Sm) and resumption (INH5Sm). Data are means SD.

ϭ otine in blood) and decreased thereafter Glucodynamic results INH1Sm–to–AUC-GLU0–360 ratio Ͼ ⌬ from 100 ng/ml at baseline to near non- The glucodynamic results were consistent 1.16; INH1Sm/INH1NSm–to– GLUmax smoking concentrations within 3 days with those seen for pharmacokinetics. ratio ϭ 1.19) with unchanged median (Table 1). Resumption of smoking caused The glucodynamic response to inhaled in- t⌬GLUmax at 53 min. The response de- restoration to baseline concentrations. sulin was much greater in smokers than in clined after cessation for 3 days and was about equivalent to that in nonsmokers nonsmokers (INH1Sm/INH1NSm–to– AUC-GLU ratio ϭ 1.73, P Ͻ 0.05; after 1 week (INH4Sm/INH1Sm–to–AUC- 0–360 ϭ Relative bioavailability INH1 /INH1 –to–⌬GLU ratio ϭ GLU0–360 ratio 0.74; INH1Sm/ Sm NSm max ⌬ ϭ Ͻ INH1NSm–to– GLUmax ratio 0.49) Bioavailability, as determined from the 2.74, P 0.001) and t⌬GLUmax was nom- with median t⌬ delayed to 82.5 AUC-Insulin0–360, of inhaled insulin rel- inally shorter in smokers, which ap- GLUmax ative to subcutaneous insulin at baseline proached signiﬁcance (median 45 vs. 75 min. The glucodynamic response was re- was 8% in nonsmokers and 12% in smok- min). Subcutaneous insulin, on the other stored to baseline after resumption of smoking for 2–3 days (INH4Sm/INH1Sm– ers. Short-term (12-h) cessation of smok- hand, was equivalent in smokers and ϭ to–AUC-GLU0–360 ratio 0.87; INH1Sm/ ing increased the bioavailability to 15% in nonsmokers (SCSm/SCNSm–to–AUC- ⌬ ϭ ϭ INH1NSm–to– GLUmax ratio 0.82) smokers. However, longer cessation of GLU0–360 ratio 1.22; SCSm/SCNSm–to– ⌬ ϭ with median t⌬GLUmax down to 56 min. smoking progressively reduced bioavail- GLUmax ratio 1.30; median t⌬GLUmax ability to 11% after 3 days and to 9% after ϭ 120 vs. 180 min). After smoking cessa- 1 week, approaching the bioavailability of tion for 12 h, the glucodynamic response Safety and tolerability 8% in nonsmokers. Resumption of smok- with inhaled insulin was not signiﬁcantly A total of four treatment-related adverse ing for 2–3 days restored the bioavailabil- altered relative to baseline smoker levels, events were reported in two subjects dur- ity back to the smoker level of 15%. although it nominally increased (INH2Sm/ ing the study (two moderate hypoglyce-

280 DIABETES CARE, VOLUME 29, NUMBER 2, FEBRUARY 2006 Becker and Associates mia, one mild headache, and one The study by Himmelman et al. (16) Diabetes health organizations unani- moderate asthenia); all were in the used a different insulin inhalation system mously recommend that people with dia- smoker group. Both cases of hypoglyce- that utilizes a liquid aerosol formulation, betes should not start or continue to mia occurred with inhaled insulin ϳ1h as opposed to the dry-powder formula- smoke, as smoking substantially increases after INH2Sm, and no action was required. tion in the present study, and also showed the risk of developing cardiovascular dis- No subject was discontinued or had a a higher AUC-Insulin0–360 and Cmax and ease and diabetes complications, espe- dose reduction owing to adverse events. a shorter tmax for inhaled insulin in smok- cially neuropathy and nephropathy (20– Spirometry tests revealed no clinically rel- ing subjects. Thus, the effects of smoking 23). The precise metabolic effects of evant changes in either group. There were on inhaled insulin absorption are not for- nicotine in people with diabetes are yet to no clinically relevant changes in vital mulation specific. Rather, they reflect a be firmly established, although some evi- signs in either group, and physical exam- general increase in airway epithelial per- dence suggests that nicotine increases in- inations also showed no treatment- meability associated with smoking, con- sulin resistance and has adverse effects on related changes. sistent with predictions from radioaerosol glucose homeostasis (21,24). Unfortu- lung clearance (8–11). In fact, the time nately, the prevalence of smoking among course observed in our study closely mir- individuals with diabetes remains compa- CONCLUSIONS — The results from rors that observed for radioaerosols with rable with the general population (20). this study are consistent with previous re- smoking cessation and resumption The results of the present study confirm ports that the absorption of inhaled insu- (9,18). For perspective, acute passive ex- that patients are required to abstain from lin and other small molecules is increased posure to tobacco smoke, in contrast to in smokers compared with nonsmokers active smoking, is expected not to alter (or smoking before starting and when con- (8–11,13–16). They are also consistent at most to reduce to a small extent) ab- tinuing treatment with inhaled insulin. with a preliminary report showing that sorption of inhaled insulin based on stud- In summary, this study shows that 1) smoking cessation for 3 weeks reduced ies with diethylenetriamine penta-acetic inhaled insulin absorption is increased in inhaled insulin absorption by ϳ50% to- acid (26). The exact mechanism underly- smokers compared with nonsmokers, ward the levels seen in nonsmokers, an ing the increased airway permeability consistent with previous studies; 2) the effect that was not reduced further after 3 with smoking is not yet fully understood, effect of smoking on insulin absorption is months’ smoking cessation (15). How- although it is nicotine-independent (12) reversed after only 1 week of smoking ces- ever, this is the first study to describe the and may be associated with increased ox- sation, although the absorption profile re- early time course of the onset of smoking idative stress (18). mains altered relative to that of cessation effects on inhaled insulin ab- In a previous study, smokers had a nonsmokers; 3) changes in absorption oc- sorption and reveals that smoking cessa- higher incidence of hypoglycemia after cur rapidly after only 12 h of smoking tion for only 1 week decreases total inhaled insulin administration, reflecting cessation and may involve a transient in- pulmonary absorption to almost that of the enhanced rate and extent of absorp- crease in absorption; and 4) the insulin healthy nonsmoker levels. However, in tion of inhaled insulin in smokers (15). In absorption profile rapidly reverts back to spite of this, the other inhaled insulin the present study, there were two inci- that seen in chronic smokers after only pharmacokinetic parameters (Cmax and dences of hypoglycemia, both of which 1–2 days of smoking resumption. We tmax) indicate that the apparent rate of ab- occurred in the smoker group with in- therefore conclude that, as short-term sorption remains altered to some extent haled insulin administration 12 h after changes in insulin availability may in- after a week of smoking cessation. This is smoking cessation. It is interesting to note crease hypoglycemia risk, people with di- also the first study to show that resump- that this is the point at which a transient abetes who smoke must quit smoking tion of smoking restores the changes that increase in insulin exposure occurred, before commencing treatment with in- occurred during smoking cessation back such that bioavailability increased from haled insulin and that inhaled insulin to the levels seen in chronic smokers after 12 to 15%. Whether this transient in- should not be used in those who choose only 2–3 days. This restoration may start crease is real or artifactual remains unde- to continue smoking. This conclusion is to occur very rapidly upon resumption of termined. However, although not consistent with recommendations that smoking; preliminary evidence from a re- designed to investigate the time course of people with diabetes should refrain from cent study reveals a tendency for pharma- smoking cessation, the study by Himmel- smoking altogether (20–23). cokinetic and glucodynamic parameters man et al. (16) appears to show a compat- to revert to precessation levels Ͻ90 min ible effect. Inhaled insulin absorption was after resumption of smoking (19). increased to a greater degree (relative to The pharmacokinetic profile of in- nonsmoking subjects) when smoking Acknowledgments— This study was pre- haled insulin reported here is consistent subjects completely abstained from sented at the Scientific Sessions of the Ameri- with previous reports for this and other smoking overnight compared with when can Diabetes Association, New Orleans, formulations (1,25). In a glucose-clamp they smoked acutely during the 30 min Louisiana, 13–19 June 2003, and the Congress of the International Diabetes Federation, study, Rave et al. (1) reported a tmax of 55 before inhaled insulin administration. 24–29 August 2003. It was supported by min using Exubera inhaled insulin, iden- Thus, it appears that smoking may also Pfizer (New York, NY) and Aventis, a member tical to that shown here. Furthermore, the have acute inhibitory effects on airway of the sanofi-aventis Group (Bridgewater, NJ), bioavailability of inhaled insulin relative permeability that are masked by the who are developing Exubera with Nektar to subcutaneous regular human insulin in chronic effects of increased permeability, Therapeutics (San Carlos, CA). our study was 8% in nonsmokers, which and produce a rapid and transient re- Richard Piechatzek and colleagues con- compares well with the 9% reported pre- bound in permeability during the 1st day ducted the study at the Go¨rlitz clinical unit of viously (1). of smoking cessation. IMFORM, Darmstadt, Germany.

DIABETES CARE, VOLUME 29, NUMBER 2, FEBRUARY 2006 281 Smoking and inhaled insulin absorption

greater perceived acceptance of insulin Diabetes Care 26:677–682, 2003 References therapy in patients with type 2 diabetes. 17. Minty BD, Jordan C, Jones JG: Rapid im- 1. Rave K, Bott S, Heinemann L, Sha S, Diabetes Care 28:427–428, 2005 provement in abnormal pulmonary epi- Becker RHA, Willavize SA, Heise T: 8. Jones JG, Minty BD, Lawler P, Hulands G, thelial permeability after stopping Time–action profile of inhaled insulin in Crawley JC, Veall N: Increased alveolar cigarettes. BMJ (Clin Res Ed) 282:1183– comparison with subcutaneously-in- epithelial permeability in cigarette smok- 1186, 1981 jected insulin lispro and regular human ers. Lancet 1:66–68, 1980 18. Morrison D, Rahman I, Lannan S, Mac- insulin. Diabetes Care 28:1077–1082, 9. Mason GR, Uszler IM, Effros RM, Reid E: Nee W: Epithelial permeability, inflam- 2005 Rapidly reversible alterations of pulmo- mation, and oxidant stress in the air 2. Hollander PA, Blonde L, Rowe R, Mehta nary epithelial permeability induced by spaces of smokers. Am J Respir Crit Care AE, Milburn JL, Hershon KS, Chiasson smoking. Chest 83:6–11, 1983 Med 159:473–479, 1999 J-L, Levin SR: Efficacy and safety of in- 10. Kennedy SM, Elwood RK, Wiggs BJ, Pare 19. Pan A, Chan C, Loh MT, Yeo KP, Much- haled insulin (Exubera) compared with PD, Hogg JC: Increased airway mucosal subcutaneous insulin therapy in patients more DB, Silverman B, De La Pena A, permeability of smokers: relationship to Wise SD: Effects of smoking (immediate with type 2 diabetes mellitus: results of a airway reactivity. Am Rev Respir Dis 129: 6-month, randomized, comparative trial. and delayed), abstinence, nicotine replace- 143–148, 1984 ment therapy and smoking resumption on Diabetes Care 27:2356–2362, 2004 11. Nolop KB, Maxwell DL, Fleming JS, 3. Quattrin T, Be´langer A, Bohannon NJV, pulmonary insulin pharmacokinetics and Braude S, Hughes JM, Royston D: A com- glucodynamics (Abstract). Diabetes 53 Schwartz SL: Efficacy and safety of in- parison of 99mTc-DTPA and 113mIn- haled insulin (Exubera) compared with (Suppl. 2):A112, 2004 DTPA aerosol clearances in humans: 20. Haire-Joshu D, Glasgow RE, Tibbs TL: subcutaneous insulin therapy in patients effects of smoking, hyperinflation, and in Smoking and diabetes. Diabetes Care 22: with type 1 diabetes: results of a 6-month, vitro oxidation. Am Rev Respir Dis 136: 1887–1898, 1999 randomized, comparative trial. Diabetes 1112–1116, 1987 21. Eliasson B: Cigarette smoking and diabe- Care 27:2622–2627, 2004 12. Minty BD, Royston D, Jones JG, Hulands tes. Prog Cardiovasc Dis 45:405–413, 4. DeFronzo RA, Bergenstal RM, Cefalu WT, GH: The effect of nicotine on pulmonary 2003 Pullman J, Lerman S, Bode BW, Phillips epithelial permeability in man. Chest 86: LS: Efficacy of inhaled insulin in patients 72–74, 1984 22. International Diabetes Federation: Diabe- with type 2 diabetes not controlled with 13. Ko¨hler D, Schlu¨ ter KJ, Kerp L, Matthys H: tes and tobacco use: a harmful combina- diet and exercise: a 12-week, randomized, Nicht radioaktives verfahren zur messung tion [article online], 2003. Available from comparative trial. Diabetes Care 28:1922– der lungenpermeabilita¨t: inhalation von http://www.idf.org. Accessed 8 March 2005 1928, 2005 insulin. Atemw-Lungenkrkh 13:S230– 23. American Diabetes Association: Smoking 5. Skyler JS, Weinstock RS, Raskin P, Yale S232, 1987 and diabetes (Position Statement). Diabe- J-F, Barrett E, Gerich JE, Gerstein HC: Use 14. Wise SD, Sathirakul K, Yeo KP, Chien JY; tes Care 27(Suppl. 1):S74–S75, 2004 of inhaled insulin in a basal/bolus insulin Aftring RP: Smoking increases the bioavail- 24. Axelsson T, Jansson P-A, Smith U, Elias- regimen in type 1 diabetic subjects: a ability of inhaled insulin, but relative insulin son B: Nicotine infusion acutely impairs 6-month, randomized, comparative trial. resistance ameliorates differences in action insulin sensitivity in type 2 diabetic pa- Diabetes Care 28:1630–1635, 2005 (Abstract). Diabetologia 44 (Suppl. 1):A5, tients but not in healthy subjects. J Intern 6. Rosenstock J, Cappelleri JC, Bolinder B, 2001 Med 249:539–544, 2001 Gerber RA: Patient satisfaction and glyce- 15. Sha S, Becker RHA, Willavise SA, Schu- 25. Patton JS, Bukar JG, Eldon MA: Clinical mic control after 1 year with inhaled in- macher DA, Lee JD, Carroll RS, Fryburg pharmacokinetics and pharmacodynam- sulin (Exubera) in patients with type 1 or DA: The effect of smoking cessation on ics of inhaled insulin. Clin Pharmacokinet type 2 diabetes. Diabetes Care 27:1318– the absorption of inhaled insulin (Exu- 43:781–801, 2004 1323, 2004 bera) (Abstract). Diabetes 51 (Suppl. 2): 26. Yates DH, Havill K, Thompson MM, Rit- 7. Freemantle N, Blonde L, Duhot D, A538, 2002 tano AB, Chu J, Glanville AR: Sidestream Hompesch M, Eggertsen R, Hobbs FD, 16. Himmelmann A, Jendle J, Mellen A, Pe- smoke inhalation decreases respiratory Martinez L, Ross S, Bolinder B, Stridde E: tersen AH, Dahl UL, Wollmer P: The im- clearance of 99mTc-DTPA acutely. Aust Availability of inhaled insulin promotes pact of smoking on inhaled insulin. NZ J Med 26:513–518, 1996

282 DIABETES CARE, VOLUME 29, NUMBER 2, FEBRUARY 2006