Ars.els-cdn.com
Examination of the effects of low-dose intraoperative fentanyl on postoperative respiratory complication rate. A pre-specified, retrospective analysis- SUPPLEMENTAL DIGITAL CONTENT - TOC \o "1-3" \h \z \uSupplemental section 1. Data sources PAGEREF _Toc1945156 \h 3Supplemental section 2. Evaluation of the confounder model PAGEREF _Toc1945157 \h 4Multivariable fractional polynomial modelling PAGEREF _Toc1945158 \h 4Choice of confounder format PAGEREF _Toc1945159 \h 4Supplemental section 3. Exploratory analyses PAGEREF _Toc1945160 \h 5Supplemental section 4. Interaction term analyses PAGEREF _Toc1945161 \h 6Supplemental section 5. Sensitivity analyses PAGEREF _Toc1945162 \h 7Preoperative opioid prescription PAGEREF _Toc1945163 \h 7Intraoperative long-acting opioid dose PAGEREF _Toc1945164 \h 7Postoperative pain management PAGEREF _Toc1945165 \h 7Propensity-score matching PAGEREF _Toc1945166 \h 8Duration of surgery PAGEREF _Toc1945167 \h 9Surgical complexity PAGEREF _Toc1945168 \h 9Individual anaesthesiologist preferences PAGEREF _Toc1945169 \h 9Year of surgery PAGEREF _Toc1945170 \h 9Exclusion of multiple procedures within four weeks PAGEREF _Toc1945171 \h 10Prespecified risk groups PAGEREF _Toc1945172 \h 10Multiple imputation of missing data PAGEREF _Toc1945173 \h 10Supplemental section 6. Supplemental tables PAGEREF _Toc1945174 \h 11Table S1: Definition of outcome PAGEREF _Toc1945175 \h 11Table S2. Association of intraoperative fentanyl dose and postoperative respiratory complications across fentanyl-dose deciles PAGEREF _Toc1945176 \h 13Table S3. Results of exploratory outcomes across fentanyl dose quartiles PAGEREF _Toc1945177 \h 14Table S4. Sensitivity analyses PAGEREF _Toc1945178 \h 16Table S5. Interaction analysis - Body Mass Index PAGEREF _Toc1945179 \h 19Table S6. Characteristics of study population by intraoperative fentanyl dose – after propensity matching of 1st and 4th quartile PAGEREF _Toc1945180 \h 20Supplemental section 7. Supplemental Figures PAGEREF _Toc1945181 \h 22Figure S1a. Area under the receiver operating characteristics curve for the primary regression model independent of fentanyl dose PAGEREF _Toc1945182 \h 22Figure S1b. Reliability plot for the primary regression model PAGEREF _Toc1945183 \h 22Figure S2. Composite and individual respiratory complications by year of surgery PAGEREF _Toc1945184 \h 23References PAGEREF _Toc1945185 \h 24Supplemental section 1. Data sourcesWe retrieved data from three different databases. From the Anaesthesia Information Management System (AIMS), we collected information regarding intraoperative management as well as baseline patient characteristics. Data on preoperative medications, comorbidities, and postoperative outcomes were collected from the Research Patient Data Registry (RPDR), a centralized registry dedicated to providing data for research purposes. From 2007 to 2014, comorbidities and outcomes were coded according to the International Classification of Diseases and Related Health Problems, Ninth Revision (ICD9), while we used the Tenth Revision (ICD10) for diagnostic codes from the beginning of 2015. Additionally, we obtained information on hospital admission and discharge as well as hospital length of stay and hospital costs from Enterprise Performance Systems Inc. (EPSi), a system for performance improvement and financial planning.Supplemental section 2. Evaluation of the confounder modelWe tested all continuous variables in the primary regression model for linearity. Due to non-linearity of the coefficients, the following variables were categorized into equally-sized quintiles: intraoperative vasopressor dose, duration of surgery, intraoperative fluid volume, work relative value units and inspiratory O2-fraction. The following confounder variables were categorized according to clinically reasonable cut-off points: body mass index (BMI) ADDIN EN.CITE <EndNote><Cite><Author>Organization</Author><Year>2000</Year><RecNum>243</RecNum><DisplayText><style face="superscript">1</style></DisplayText><record><rec-number>243</rec-number><foreign-keys><key app="EN" db-id="f0d0fwr06ptsdseatwtv5a2t9pewsvav50px" timestamp="1532377018">243</key></foreign-keys><ref-type name="Report">27</ref-type><contributors><authors><author>World Health Organization</author></authors></contributors><titles><title>Obesity: preventing and managing the global epidemic
Report of a WHO Consultation </title><secondary-title>WHO Technical Report Series </secondary-title></titles><pages>252</pages><number>ISBN: 92 4 120894 5</number><num-vols>894</num-vols><dates><year>2000</year></dates><urls></urls></record></Cite></EndNote>1, American Society of Anesthesiologists (ASA) physical status classification (0-2, 3-5), Score for Prediction of Postoperative Respiratory Complication (SPORC),PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5CcnVlY2ttYW5uPC9BdXRob3I+PFllYXI+MjAxMzwvWWVh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=
ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5CcnVlY2ttYW5uPC9BdXRob3I+PFllYXI+MjAxMzwvWWVh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=
ADDIN EN.CITE.DATA 2 Score for Prediction of Obstructive Sleep Apnoea (SPOSA),PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5TaGluPC9BdXRob3I+PFllYXI+MjAxNzwvWWVhcj48UmVj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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5TaGluPC9BdXRob3I+PFllYXI+MjAxNzwvWWVhcj48UmVj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ADDIN EN.CITE.DATA 3 and intraoperative hypotensive minutes (0-10, 11-19, ≥ 20) as well as units of packed red blood cells (0, 1, 2, ≥ 3).Multivariable fractional polynomial modellingIn our primary model, five covariates not underlying the linearity assumption were categorized into equally sized quintiles. In-order to adjust for potential insufficient consideration of non-linearity in these continuous covariates, we performed a sensitivity analysis utilizing multivariable fractional polynomial modelling.PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5Sb3lzdG9uPC9BdXRob3I+PFllYXI+MTk5OTwvWWVhcj48
UmVjTnVtPjI3OTwvUmVjTnVtPjxEaXNwbGF5VGV4dD48c3R5bGUgZmFjZT0ic3VwZXJzY3JpcHQi
PjQ8L3N0eWxlPjwvRGlzcGxheVRleHQ+PHJlY29yZD48cmVjLW51bWJlcj4yNzk8L3JlYy1udW1i
ZXI+PGZvcmVpZ24ta2V5cz48a2V5IGFwcD0iRU4iIGRiLWlkPSJmMGQwZndyMDZwdHNkc2VhdHd0
djVhMnQ5cGV3c3ZhdjUwcHgiIHRpbWVzdGFtcD0iMTU1MDE4NzM2NSI+Mjc5PC9rZXk+PC9mb3Jl
aWduLWtleXM+PHJlZi10eXBlIG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10eXBlPjxj
b250cmlidXRvcnM+PGF1dGhvcnM+PGF1dGhvcj5Sb3lzdG9uLCBQLjwvYXV0aG9yPjxhdXRob3I+
QW1ibGVyLCBHLjwvYXV0aG9yPjxhdXRob3I+U2F1ZXJicmVpLCBXLjwvYXV0aG9yPjwvYXV0aG9y
cz48L2NvbnRyaWJ1dG9ycz48YXV0aC1hZGRyZXNzPkRlcGFydG1lbnQgb2YgTWVkaWNhbCBTdGF0
aXN0aWNzICZhbXA7IEV2YWx1YXRpb24sIEltcGVyaWFsIENvbGxlZ2UgU2Nob29sIG9mIE1lZGlj
aW5lLCBMb25kb24sIFVLLiBwcm95c3RvbkBpYy5hYy51azwvYXV0aC1hZGRyZXNzPjx0aXRsZXM+
PHRpdGxlPlRoZSB1c2Ugb2YgZnJhY3Rpb25hbCBwb2x5bm9taWFscyB0byBtb2RlbCBjb250aW51
b3VzIHJpc2sgdmFyaWFibGVzIGluIGVwaWRlbWlvbG9neTwvdGl0bGU+PHNlY29uZGFyeS10aXRs
ZT5JbnQgSiBFcGlkZW1pb2w8L3NlY29uZGFyeS10aXRsZT48YWx0LXRpdGxlPkludGVybmF0aW9u
YWwgam91cm5hbCBvZiBlcGlkZW1pb2xvZ3k8L2FsdC10aXRsZT48L3RpdGxlcz48cGVyaW9kaWNh
bD48ZnVsbC10aXRsZT5JbnQgSiBFcGlkZW1pb2w8L2Z1bGwtdGl0bGU+PGFiYnItMT5JbnRlcm5h
dGlvbmFsIGpvdXJuYWwgb2YgZXBpZGVtaW9sb2d5PC9hYmJyLTE+PC9wZXJpb2RpY2FsPjxhbHQt
cGVyaW9kaWNhbD48ZnVsbC10aXRsZT5JbnQgSiBFcGlkZW1pb2w8L2Z1bGwtdGl0bGU+PGFiYnIt
MT5JbnRlcm5hdGlvbmFsIGpvdXJuYWwgb2YgZXBpZGVtaW9sb2d5PC9hYmJyLTE+PC9hbHQtcGVy
aW9kaWNhbD48cGFnZXM+OTY0LTc0PC9wYWdlcz48dm9sdW1lPjI4PC92b2x1bWU+PG51bWJlcj41
PC9udW1iZXI+PGVkaXRpb24+MTk5OS8xMi8yMjwvZWRpdGlvbj48a2V5d29yZHM+PGtleXdvcmQ+
Q29yb25hcnkgRGlzZWFzZS9lcGlkZW1pb2xvZ3kvKm1vcnRhbGl0eTwva2V5d29yZD48a2V5d29y
ZD5Dcm9zcy1TZWN0aW9uYWwgU3R1ZGllczwva2V5d29yZD48a2V5d29yZD4qRXBpZGVtaW9sb2dp
YyBNZXRob2RzPC9rZXl3b3JkPjxrZXl3b3JkPkh1bWFuczwva2V5d29yZD48a2V5d29yZD5MaW5l
YXIgTW9kZWxzPC9rZXl3b3JkPjxrZXl3b3JkPk1hbGU8L2tleXdvcmQ+PGtleXdvcmQ+Kk1vZGVs
cywgQ2FyZGlvdmFzY3VsYXI8L2tleXdvcmQ+PGtleXdvcmQ+TW9ydGFsaXR5Lyp0cmVuZHM8L2tl
eXdvcmQ+PGtleXdvcmQ+TmVvcGxhc21zL2VwaWRlbWlvbG9neS8qbW9ydGFsaXR5PC9rZXl3b3Jk
PjxrZXl3b3JkPk9kZHMgUmF0aW88L2tleXdvcmQ+PGtleXdvcmQ+UHJvc3BlY3RpdmUgU3R1ZGll
czwva2V5d29yZD48a2V5d29yZD5SaXNrIEFzc2Vzc21lbnQ8L2tleXdvcmQ+PGtleXdvcmQ+Umlz
ayBGYWN0b3JzPC9rZXl3b3JkPjxrZXl3b3JkPlNlbnNpdGl2aXR5IGFuZCBTcGVjaWZpY2l0eTwv
a2V5d29yZD48a2V5d29yZD5TdHJva2UvZXBpZGVtaW9sb2d5Lyptb3J0YWxpdHk8L2tleXdvcmQ+
PGtleXdvcmQ+VW5pdGVkIEtpbmdkb20vZXBpZGVtaW9sb2d5PC9rZXl3b3JkPjwva2V5d29yZHM+
PGRhdGVzPjx5ZWFyPjE5OTk8L3llYXI+PHB1Yi1kYXRlcz48ZGF0ZT5PY3Q8L2RhdGU+PC9wdWIt
ZGF0ZXM+PC9kYXRlcz48aXNibj4wMzAwLTU3NzEgKFByaW50KSYjeEQ7MDMwMC01NzcxPC9pc2Ju
PjxhY2Nlc3Npb24tbnVtPjEwNTk3OTk4PC9hY2Nlc3Npb24tbnVtPjx1cmxzPjwvdXJscz48cmVt
b3RlLWRhdGFiYXNlLXByb3ZpZGVyPk5MTTwvcmVtb3RlLWRhdGFiYXNlLXByb3ZpZGVyPjxsYW5n
dWFnZT5lbmc8L2xhbmd1YWdlPjwvcmVjb3JkPjwvQ2l0ZT48L0VuZE5vdGU+AG==
ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5Sb3lzdG9uPC9BdXRob3I+PFllYXI+MTk5OTwvWWVhcj48
UmVjTnVtPjI3OTwvUmVjTnVtPjxEaXNwbGF5VGV4dD48c3R5bGUgZmFjZT0ic3VwZXJzY3JpcHQi
PjQ8L3N0eWxlPjwvRGlzcGxheVRleHQ+PHJlY29yZD48cmVjLW51bWJlcj4yNzk8L3JlYy1udW1i
ZXI+PGZvcmVpZ24ta2V5cz48a2V5IGFwcD0iRU4iIGRiLWlkPSJmMGQwZndyMDZwdHNkc2VhdHd0
djVhMnQ5cGV3c3ZhdjUwcHgiIHRpbWVzdGFtcD0iMTU1MDE4NzM2NSI+Mjc5PC9rZXk+PC9mb3Jl
aWduLWtleXM+PHJlZi10eXBlIG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10eXBlPjxj
b250cmlidXRvcnM+PGF1dGhvcnM+PGF1dGhvcj5Sb3lzdG9uLCBQLjwvYXV0aG9yPjxhdXRob3I+
QW1ibGVyLCBHLjwvYXV0aG9yPjxhdXRob3I+U2F1ZXJicmVpLCBXLjwvYXV0aG9yPjwvYXV0aG9y
cz48L2NvbnRyaWJ1dG9ycz48YXV0aC1hZGRyZXNzPkRlcGFydG1lbnQgb2YgTWVkaWNhbCBTdGF0
aXN0aWNzICZhbXA7IEV2YWx1YXRpb24sIEltcGVyaWFsIENvbGxlZ2UgU2Nob29sIG9mIE1lZGlj
aW5lLCBMb25kb24sIFVLLiBwcm95c3RvbkBpYy5hYy51azwvYXV0aC1hZGRyZXNzPjx0aXRsZXM+
PHRpdGxlPlRoZSB1c2Ugb2YgZnJhY3Rpb25hbCBwb2x5bm9taWFscyB0byBtb2RlbCBjb250aW51
b3VzIHJpc2sgdmFyaWFibGVzIGluIGVwaWRlbWlvbG9neTwvdGl0bGU+PHNlY29uZGFyeS10aXRs
ZT5JbnQgSiBFcGlkZW1pb2w8L3NlY29uZGFyeS10aXRsZT48YWx0LXRpdGxlPkludGVybmF0aW9u
YWwgam91cm5hbCBvZiBlcGlkZW1pb2xvZ3k8L2FsdC10aXRsZT48L3RpdGxlcz48cGVyaW9kaWNh
bD48ZnVsbC10aXRsZT5JbnQgSiBFcGlkZW1pb2w8L2Z1bGwtdGl0bGU+PGFiYnItMT5JbnRlcm5h
dGlvbmFsIGpvdXJuYWwgb2YgZXBpZGVtaW9sb2d5PC9hYmJyLTE+PC9wZXJpb2RpY2FsPjxhbHQt
cGVyaW9kaWNhbD48ZnVsbC10aXRsZT5JbnQgSiBFcGlkZW1pb2w8L2Z1bGwtdGl0bGU+PGFiYnIt
MT5JbnRlcm5hdGlvbmFsIGpvdXJuYWwgb2YgZXBpZGVtaW9sb2d5PC9hYmJyLTE+PC9hbHQtcGVy
aW9kaWNhbD48cGFnZXM+OTY0LTc0PC9wYWdlcz48dm9sdW1lPjI4PC92b2x1bWU+PG51bWJlcj41
PC9udW1iZXI+PGVkaXRpb24+MTk5OS8xMi8yMjwvZWRpdGlvbj48a2V5d29yZHM+PGtleXdvcmQ+
Q29yb25hcnkgRGlzZWFzZS9lcGlkZW1pb2xvZ3kvKm1vcnRhbGl0eTwva2V5d29yZD48a2V5d29y
ZD5Dcm9zcy1TZWN0aW9uYWwgU3R1ZGllczwva2V5d29yZD48a2V5d29yZD4qRXBpZGVtaW9sb2dp
YyBNZXRob2RzPC9rZXl3b3JkPjxrZXl3b3JkPkh1bWFuczwva2V5d29yZD48a2V5d29yZD5MaW5l
YXIgTW9kZWxzPC9rZXl3b3JkPjxrZXl3b3JkPk1hbGU8L2tleXdvcmQ+PGtleXdvcmQ+Kk1vZGVs
cywgQ2FyZGlvdmFzY3VsYXI8L2tleXdvcmQ+PGtleXdvcmQ+TW9ydGFsaXR5Lyp0cmVuZHM8L2tl
eXdvcmQ+PGtleXdvcmQ+TmVvcGxhc21zL2VwaWRlbWlvbG9neS8qbW9ydGFsaXR5PC9rZXl3b3Jk
PjxrZXl3b3JkPk9kZHMgUmF0aW88L2tleXdvcmQ+PGtleXdvcmQ+UHJvc3BlY3RpdmUgU3R1ZGll
czwva2V5d29yZD48a2V5d29yZD5SaXNrIEFzc2Vzc21lbnQ8L2tleXdvcmQ+PGtleXdvcmQ+Umlz
ayBGYWN0b3JzPC9rZXl3b3JkPjxrZXl3b3JkPlNlbnNpdGl2aXR5IGFuZCBTcGVjaWZpY2l0eTwv
a2V5d29yZD48a2V5d29yZD5TdHJva2UvZXBpZGVtaW9sb2d5Lyptb3J0YWxpdHk8L2tleXdvcmQ+
PGtleXdvcmQ+VW5pdGVkIEtpbmdkb20vZXBpZGVtaW9sb2d5PC9rZXl3b3JkPjwva2V5d29yZHM+
PGRhdGVzPjx5ZWFyPjE5OTk8L3llYXI+PHB1Yi1kYXRlcz48ZGF0ZT5PY3Q8L2RhdGU+PC9wdWIt
ZGF0ZXM+PC9kYXRlcz48aXNibj4wMzAwLTU3NzEgKFByaW50KSYjeEQ7MDMwMC01NzcxPC9pc2Ju
PjxhY2Nlc3Npb24tbnVtPjEwNTk3OTk4PC9hY2Nlc3Npb24tbnVtPjx1cmxzPjwvdXJscz48cmVt
b3RlLWRhdGFiYXNlLXByb3ZpZGVyPk5MTTwvcmVtb3RlLWRhdGFiYXNlLXByb3ZpZGVyPjxsYW5n
dWFnZT5lbmc8L2xhbmd1YWdlPjwvcmVjb3JkPjwvQ2l0ZT48L0VuZE5vdGU+AG==
ADDIN EN.CITE.DATA 4In the resulting model, intraoperative vasopressor dose was treated as polynomials with the power -2, duration of surgery with the power -1 and intraoperative fluid volume with -0.5 accordingly, while for the variable work relative value units fractional polynomials with the power of 0.5 were determined and of 2 for inspiratory O2-fraction.The sensitivity analysis utilizing multivariable fractional polynomial modelling confirmed our primary findings: a significant association of low intraoperative fentanyl dose and lower odds of PRCs (Q1 vs Q4: aOR 0.80, CI 0.76-0.84, P<0.001; Table S4 F).Choice of confounder formatIn the primary model, dose of volatile anaesthetics was expressed as average MAC (minimum alveolar concentration at 1 atmosphere that prevents movement in 50% of patients exposed to a surgical incision) of the case as calculated from the mean end-tidal concentration of the respective gas throughout the case and was adjusted by an anaesthetic-specific age-factor. Fraction of inspired oxygen (FiO2) was defined as the median FiO2 during the case. In response to concerns in peer-review, a sensitivity analysis was performed utilizing the format MAC-hours and FiO2-hours to account for dose of volatile anaesthetics and fraction of inspired oxygen over the course of the procedure. Utilizing alternative methods to express the confounders dose of volatile anaesthetics and fraction of inspired oxygen, demonstrated the robustness of our findings: Q1 vs Q4 - aOR 0.81, CI 0.77-0.86, P<0.001.Supplemental section 3. Exploratory analysesWith an exploratory intent, we evaluated the association between fentanyl dose and length of hospital stay, total costs of hospital care, incidence of surgical site infection within 30 days after surgery, postoperative ICU admission and perioperative naloxone use as well as post-extubation desaturation, mortality within 30 days after surgery, non-invasive ventilation, constipation and delirium within 3 days after surgery. Costs were defined as the sum of direct costs, based on patient services, and indirect costs, the latter representing costs not directly linked to patient care (e.g. costs from the financial services department). Low intraoperative fentanyl dose was significantly associated with a lower risk of ICU admission during the index hospital stay, surgical site infection within 30 days after surgery and lower perioperative naloxone use. Measures of healthcare utilization (i.e., hospital costs and length of hospital stay) were significantly lower among patients receiving low intraoperative fentanyl doses (Table S3).Low fentanyl dose was significantly associated with a lower risk of post-extubation desaturation below 80%, but not below 90%. There was no significant association between intraoperative fentanyl dose and risk of non-invasive ventilation, constipation or delirium within three days after surgery or mortality within 30 days after surgery (Table S3).Supplemental section 4. Interaction term analysesMethodsTo identify vulnerable subcohorts of patients, we examined interaction terms concerning patient characteristics (“fentanyl dose x age”, “fentanyl dose x BMI”), surgery location (“fentanyl dose x thoracic surgery”, “fentanyl dose x abdominal surgery”), and other intraoperative medications (“fentanyl dose x mean alveolar concentration [MAC] of inhalational anaesthetics”, “fentanyl dose x neuromuscular blocking agents [NMBA] dose”) by integrating the interaction terms separately into the primary model. In order to determine clinically meaningful differences in risk, we estimated the adjusted absolute risk difference (aARD) between low and high fentanyl dose (Q1 vs Q4), i.e. the risk of PRCs attributable to intraoperative fentanyl dose, in the respective subgroups with documented significant effect modification. For continuous variables, we defined equally-sized quartiles adopting the same methodology as for the exposure variable. ResultsWhile abdominal surgery did not modify the examined association (interaction term “fentanyl dose x abdominal surgery”: p for interaction = 0.81), thoracic surgery modified the association of low fentanyl dose and lower odds of PRCs compared with high fentanyl dose towards a more substantial difference in risk among patients undergoing thoracic surgery: -6.2% (Q1 vs Q4, CI -9.7 to -2.5%; P for interaction <0.001, OR 1.21, CI 1.17-1.25) (Table S4A)..Depth of anaesthesia (i.e. MAC of volatile anaesthetics) also modified the association between intraoperative fentanyl dose and PRCs significantly (interaction term “fentanyl dose x MAC”: P for interaction =0.016, OR 1.03, CI 1.01-1.06). Although the estimated risk of PRCs, independent of fentanyl dose, was lower among patients receiving high doses of volatile anaesthetics (5.1-7.3%), the difference in risk of PRCs attributable to low fentanyl dose was more substantial: -2.2% (Q1 vs Q4, CI -2.9 to -1.5%; Table S4 B). Similarly, patients receiving high doses of NMBA demonstrated a greater difference in risk of PRCs attributable to low-dose fentanyl: -3.4% (Q1 vs Q4, CI -4.6 to -2.3%; interaction term “fentanyl dose x NMBA dose”: P for interaction <0.001, OR 0.99, CI 0.99-0.99; Table S4 C). Among patients receiving high doses of both, volatile anaesthetics and NMBA, the difference in risk of PRCs attributable to low-dose fentanyl in comparison to high-dose fentanyl added up to -4.1% (Q1 vs Q4, CI -5.8 to -2.3%; Table S4 D). Age did not modify the examined association significantly (interaction term “fentanyl dose x age”: P for interaction = 0.20). The interaction term “fentanyl dose x BMI” showed significance (p for interaction < 0.001, OR 1.00, CI 1.00-1.01), but the difference in risk of PRCs attributable to low fentanyl dose of -1.5% (Q1 vs Q4, CI -2.5 to -0.4%) among severely obese patients (BMI ≥ 35 kg m-2) was not meaningfully different from that among patients with a lower BMI: -1.5% (Q1 vs Q4, CI -1.9 to -1.1%) (Table S5).Supplemental section 5. Sensitivity analysesPreoperative opioid prescriptionIn our primary model, we controlled for preoperative opioid prescription. As it is challenging to capture the actual use of opioids, in a sensitivity analysis we utilized alternative classifications to control for preoperative opioid use (categorization in accordance to World Health Organization [WHO] strength classifications ADDIN EN.CITE <EndNote><Cite><Year>1996</Year><RecNum>32</RecNum><DisplayText><style face="superscript">5</style></DisplayText><record><rec-number>32</rec-number><foreign-keys><key app="EN" db-id="vrz0fvfa2xfp0oetxr0ptptsa9spaeppftd2" timestamp="1520270768">32</key></foreign-keys><ref-type name="Report">27</ref-type><contributors></contributors><titles><title>Cancer pain relief: with a guide to opioid availability</title></titles><edition>Second</edition><dates><year>1996</year></dates><pub-location>Geneva</pub-location><publisher>World Health Organization</publisher><urls></urls></record></Cite></EndNote>5 and total number of unique preoperative opioid compounds). Additionally, we re-examined the primary hypothesis among patients without prescription of any opioid within 90 days prior to surgery.In our study cohort, 11% of all patients were prescribed any opioid medication within three months prior to surgery according to medical records. Use of two or more different opioid compounds within three months preoperatively was recorded in 3.8% of all patients and use of three or more different compounds was recorded in 1.2% of patients. According to WHO strength classification, 8.8% of patients received a prescription for a ‘strong’ opioid, 2.2% for only ‘weak’ compounds. When controlling for number of opioid compounds or for strength of prescribed preoperative opioid in the primary model, the primary findings were confirmed. Among the 129,725 patients without preoperative opioid prescription, the association of high intraoperative fentanyl dose and PRCs risk remained robust (Q1 vs Q4: aOR 0.79, CI 0.75-0.84, P<0.001; Table S4 C).Intraoperative long-acting opioid doseIn the primary model, we adjusted for intraoperative dose of long-acting opioids such as methadone, morphine, hydromorphone and sufentanil (morphine equivalent). As a sensitivity measure, we tested if intraoperative long-acting opioid dose would modify the association of high intraoperative fentanyl dose and PRCs risk by integrating the corresponding interaction term into the primary regression model.There was no significant interaction by intraoperative long-acting opioid dose: interaction term “fentanyl dose x long-acting opioid dose” – P for interaction =0.099.In addition, in order to account for residual confounding due to combination of equivalent doses of opioids with different pharmacological characteristics, we performed a sensitivity analysis among 48,769 patients who did not receive any long-acting opioid intraoperatively. Our results remained robust in this subset of patients: Q1 vs Q4 – aOR 0.87, CI 0.79-0.95, P=0.002 (Table S4 D).Postoperative pain managementIn sensitivity analysis, we a) captured the number of different opioid compounds a patient was prescribed postoperatively on the day of discharge and b) alternatively categorized the compounds based on the WHO strength classification. We then included postoperative opioid prescription categorized according to a) and b) separately in the primary model as additional confounders. Additionally, we performed a sensitivity analysis in a subgroup of patients that were prescribed any opioids postoperatively.Similarly, we also categorized patients based on prescriptions of non-opioid analgesics on discharge and performed sensitivity analyses in the following subgroups: patients that were not prescribed any pain medication (opioids or non-opioid analgesics) postoperatively and patients with an non-opioid analgesics?prescription. Of the 77,106 patients with postoperative opioid prescription, 88.3% received prescriptions for at least one ‘strong’ opioid compound and 10.8% were prescribed two or more different compounds. Adjustment for postoperative opioid prescription demonstrated the robustness of our findings: a) number of different compounds - Q1 vs Q4: aOR 0.79, CI 0.75-0.84, P<0.001; b) strength classification – Q1 vs Q4: aOR 0.79, CI 0.75-0.83, P<0.001. The association of low intraoperative fentanyl dose and lower odds of PRC remained significant among patients with a postoperative opioid prescription (n=77,106, Q1 vs Q4: aOR 0.74, CI 0.68-0.80, P<0.001, Table S4 E)Low intraoperative fentanyl dose was significantly associated with lower odds of PRC was significant among patients with prescription of a non-opioid analgesic postoperatively (n=71,570, Q1 vs Q4: aOR 0.76, CI 0.70-0.82, P<0.001, Table S4 E) as well as among the 56,675 (38.9%) patients that were not prescribed any pain medication on discharge (Q1 vs Q4: aOR 0.82, CI 0.75-0.89, P<0.001; Table S4 E). All patients underwent surgery with general anaesthesia. In our primary confounder model, we adjusted for the intraoperative use of neuraxial anaesthesia in addition to general anaesthesia. In a sensitivity analysis, we account for the intraoperative use of peripheral nerve blockage as additional confounder in our model. In this analysis the association of low intraoperative fentanyl dose and lower PRC rate remained robust: Q1 vs Q4 - aOR 0.79, CI 0.75-0.83, P<0.001 (Table S4 G).Propensity-score matchingIn order to account for unmeasured confounding, we aimed to address underlying bias leading to patients receiving high fentanyl doses with propensity-score matching. The following variables were utilized to estimate the propensity for receiving a high intraoperative fentanyl dose (fourth quartile, Q4): year of surgery, gender, age, body mass index, admission type , ASA physical status classification, Score for Prediction of Postoperative Respiratory Complication (SPORC)PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5CcnVlY2ttYW5uPC9BdXRob3I+PFllYXI+MjAxMzwvWWVh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=
ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5CcnVlY2ttYW5uPC9BdXRob3I+PFllYXI+MjAxMzwvWWVh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=
ADDIN EN.CITE.DATA 2, including high risk surgical service and emergency surgery status; Score for Prediction of Obstructive Sleep Apnoea (SPOSA)PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5TaGluPC9BdXRob3I+PFllYXI+MjAxNzwvWWVhcj48UmVj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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5TaGluPC9BdXRob3I+PFllYXI+MjAxNzwvWWVhcj48UmVj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ADDIN EN.CITE.DATA 3, including history of hypertension, atrial fibrillation, chronic pulmonary disease, congestive heart failure, diabetes, dyslipidaemia, hemi/paraplegia, liver disease, pulmonary hypertension and coronary artery disease within one year prior to surgery; code status, prescription of any opioid within 90 days prior to surgery, use of neuraxial anaesthesia, duration of surgery, work relative value units, neuromuscular blocking agent dose, vasopressor dose, dose of long-acting opioids such as methadone, morphine, hydromorphone and sufentanil, dose of volatile anaesthetics, units of packed red blood cells, intravenous fluid volume, protective ventilator settings (driving pressure ≤ 15 mmHg), inspiratory O2-fraction and intraoperative hypotension. All of these factors were forced into a logistic regression model to build the propensity score. In a sensitivity analysis, we re-examined the hypothesis that patients receiving low fentanyl doses (first quartile, Q1) would have lower odds of PRC than patients receiving high fentanyl doses (Q4) in a subset of patients matched by their propensity to receive a high intraoperative fentanyl dose. Matching was performed 1:1 with a calliper set to 0.10 using nearest neighbour and sampling without replacement. A total 39,774 of 73,107 (54.4%) patients from Q1 and Q4 were matched with a 1:1 ratio (Table S6). Re-examination of the primary finding in the matched cohort confirmed the association of low fentanyl dose and lower odds of PRC: OR 0.75, CI 0.71-0.79, P<0.001.The following variables retained residual imbalances in the matched sample and were therefore forced into the regression model: gender, body mass index, duration of surgery, work relative value units, neuromuscular blocking dose and use of neuraxial anaesthesia (Table S6). This analysis confirmed our primary results as well (aOR 0.84, CI 0.79-0.89, P<0.001).Duration of surgeryIn the primary analysis, duration of surgery was included as a covariate. Additionally, we examined the interaction term “fentanyl dose x duration of surgery” by integrating it into the primary regression model. In case of significant effect modification by duration of surgery, subsequent sensitivity analyses in subgroups of short (under two hours) and long procedural duration (≥ two hours) were performed.The interaction term “fentanyl dose x duration of surgery” demonstrated significant effect modification (p for interaction 0.005, OR 1.003, CI 1.001-1.005). Inclusion of this term in the primary regression analysis revealed a robust association of low fentanyl dose and lower PRC rates (Q1 vs Q4: aOR 0.83, CI 0.78-0.88, P<0.001; Table S4 H). Subsequent subgroup analyses of procedural duration confirmed our primary findings (Table S4 J).Surgical complexityThe following variables were included in the primary model to adjust for surgical complexity: high risk surgical service, emergency surgery status, duration of surgery, work relative value units as well as administered vasopressor dose, units of packed red blood cell units and intravenous fluid administration volume. To address remaining bias due to surgical complexity, we performed a mixed effects logistic regression analysis adjusting for the primary surgical procedure. Clustering within the surgical procedure was set as random effect while all a priori defined covariates were included as fixed effects. Information on the primary surgical procedures was available for 130,420 (89.5%) cases, including 3353 different primary procedures. Clustering within the primary surgical procedure in mixed effects analysis confirmed a robust association of low intraoperative fentanyl dose and lower odds of PRC (Q1 vs Q4: aOR 0.88, CI 0.83-0.94, P<0.001; Table S4 J).Individual anaesthesiologist preferencesWe performed mixed effects logistic regression analysis to account for differences in fentanyl use between anaesthesiologists. The random effect was clustering within anaesthesia provider and fixed effects were all predefined confounder variables.Identification of the anaesthesia provider was possible in 145,725 (99.99%) surgical cases under the care of 751 individual anaesthesiologists. The primary association remained stable when accounting for differences in fentanyl use between anaesthesiologists in mixed effects regression analysis (Q1 vs Q4: aOR 0.79, CI 0.75-0.83, P<0.001) (Table S4 K).Year of surgeryOur study encompasses non-cardiac surgical cases from 2007 to 2015. During this eight-year study period standards of care and institutional practices may have undergone significant changes. Figure S2 shows the rate of the composite outcome and the individual postoperative respiratory complications within 3 days after surgery by year of procedure. In a sensitivity analysis, we added year of surgery as a categorical variable to the primary model. Adjusting for year of surgery did not alter our primary findings and demonstrated lower odds of PRC with low fentanyl doses (Q1 vs Q4: aOR 0.77, CI 0.73-0.81, P<0.001; Table S4 L).Exclusion of multiple procedures within four weeksAs our study period spans nine years, some patients are included multiple times for index surgeries at different dates within the study period. In a sensitivity analysis, we re-examined the primary hypothesis after exclusion of cases in which the same patient had undergone surgery within four weeks prior to the index surgery.In this sensitivity analysis, we re-examined the primary hypothesis in a total of 136,696 cases. The association of low intraoperative fentanyl dose (Q1 vs Q4) and lower PRC rates was maintained: aOR 0.79 (CI 0.75-0.84, P<0.001; Table S4 M). Prespecified risk groupsTo further test the robustness of our findings, we performed subgroup analyses in different pre-specified risk groups characterized by Charlson Comorbidity Index (CCI) ADDIN EN.CITE <EndNote><Cite><Author>Charlson</Author><Year>1994</Year><RecNum>54</RecNum><DisplayText><style face="superscript">6</style></DisplayText><record><rec-number>54</rec-number><foreign-keys><key app="EN" db-id="f0d0fwr06ptsdseatwtv5a2t9pewsvav50px" timestamp="1500063205">54</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Charlson, Mary</author><author>Szatrowski, Ted P.</author><author>Peterson, Janey</author><author>Gold, Jeffrey</author></authors></contributors><titles><title>Validation of a combined comorbidity index</title><secondary-title>Journal of Clinical Epidemiology</secondary-title></titles><periodical><full-title>Journal of Clinical Epidemiology</full-title></periodical><pages>1245-1251</pages><volume>47</volume><number>11</number><keywords><keyword>Comorbidity</keyword><keyword>Prognosis</keyword></keywords><dates><year>1994</year><pub-dates><date>1994/11/01/</date></pub-dates></dates><isbn>0895-4356</isbn><urls><related-urls><url>(94)90129-5</electronic-resource-num></record></Cite></EndNote>6 (<3, ≥3), ASA status classification (0-2, 3-5), high risk vs. low risk surgery (based on surgical service)PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5CcnVlY2ttYW5uPC9BdXRob3I+PFllYXI+MjAxMzwvWWVh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=
ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5CcnVlY2ttYW5uPC9BdXRob3I+PFllYXI+MjAxMzwvWWVh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=
ADDIN EN.CITE.DATA 2 as well as patients with and without pre-existing respiratory comorbidities, defined as a diagnosis of obstructive sleep apnoea (OSA), asthma, chronic obstructive pulmonary disease (COPD), or pulmonary hypertension within one year prior to surgery.The association of low intraoperative fentanyl dose and lower odds of PRCs remained robust in all different pre-specified risk groups (Table S4 N-Q).On reviewer suggestion, an additional sensitivity analysis was performed examining the robustness of our findings in subgroups of patients undergoing ambulatory or inpatient procedures. The association of low fentanyl doses and lower odds of PRCs retained significance among both patients undergoing ambulatory procedures (Q1 vs Q4: aOR 0.68, CI 0.56-0.82, P<0.001) and inpatients (Q1 vs Q4: aOR 0.80, CI 0.76-0.85, P<0.001; Table S4 R).Multiple imputation of missing dataOur primary analysis was performed in a complete case cohort. To account for bias due to pattern of missingness, we performed multiple imputation by chained equations using STATA. After five imputations with five iterations each, the primary analysis was repeated in the imputed dataset.A total of 11 confounder variables had missing data, with SPOSA and protective ventilation being the variables missing most frequently (9,687 [5.8%] and 9,072 [5.4%] cases, respectively). Re-evaluation of the primary hypothesis in the imputed study cohort showed consistency with our findings in the complete case cohort (Q1 vs Q4: aOR 0.81, CI 0.77-0.85, P<0.001; n=166,896; Table S4 S). Supplemental section 6. Supplemental tablesTable S1: Definition of outcomeOutcome Procedure Name Code Type*Code Pneumonia Pneumococcal pneumonia [Streptococcus pneumonia] ICD-9/ICD-10 481/J13 Pneumonia due to Klebsiella pneumoniae ICD-9/ICD-10 482.0/J15.0 Pneumonia due to Pseudomonas ICD-9/ICD-10 482.1/J15.1 Pneumonia due to Streptococcus, unspecified ICD-9/ICD-10 482.30/J15.4 Pneumonia due to Staphylococcus, unspecified ICD-9/ICD-10 482.40/J15.20 Pneumonia due to Staphylococcus aureus ICD-9/ICD-10 482.41/J15.211 Methicillin resistant pneumonia due to staphylococcus aureus ICD-9/ICD-10 482.42/J15.212 Pneumonia due to Escherichia coli [E. coli] ICD-9/ICD-10 482.82/J15.5 Pneumonia due to other gram-negative bacteria ICD-9/ICD-10 482.83/J15.6 Pneumonia due to other specified bacteria ICD-9/ICD-10 482.89/J15.8 Bacterial pneumonia, unspecified ICD-9/ICD-10 482.9/ J15.9 Pneumonia, organism unspecified ICD-9/ICD-10 486/J18.9 Pneumonia due to other specified organism ICD-9/ICD-10 483.8/J16.8 Pneumonia in aspergillosis ICD-9/ICD-10 484.6/B44.0 Bronchopneumonia, organism unspecified ICD-9/ICD-10 485/J18.0 Pneumonitis due to inhalation of food or vomitus ICD-9/ICD-10 507.0/J69.0 Pulmonary Oedema Pulmonary congestion and hypostasis ICD-9/ICD-10 514/J81.1 Acute oedema of lung, unspecified ICD-9/ICD-10 518.4/J81.0 Congestive heart failure ICD-9/ICD-10 428.0/J50.9 Fluid overload ICD-9/ICD-10 276.61/E87.70 Other fluid overload ICD-9/ICD-10 276.69/E87.79 Reintubation Intubation, endotracheal, emergency procedure CPT 31500Ventilation assist and management, initiation of pressure or volume preset ventilators for assisted or controlled breathing; hospital inpatient/observation, initial day CPT 94002Respiratory Failure Pulmonary insufficiency following trauma and surgery ICD-9/ICD-10 518.5/J95.822 Acute respiratory failure following trauma and surgery ICD-9/ICD-10 518.51/J95.821 Other pulmonary insufficiency, not elsewhere classified, following trauma and surgery ICD-9/ICD-10 518.52/J95.3 Acute pulmonary insufficiency following thoracic surgery ICD-10 J95.1 Acute pulmonary insufficiency following non-thoracic surgery ICD-10 J95.2 Respiratory failure ICD-9/ICD-10 518.81/J96.90 Other pulmonary insufficiency, not elsewhere classified ICD-9/ICD-10 518.82/J80 Acute and chronic respiratory failure ICD-9/ICD-10 518.84/J96.20 AtelectasisPulmonary collapseICD-9/ICD-10518.0/J89.19AtelectasisICD-10J89.11* ICD-9: International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM); ICD-10: International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM); CPT: Current Procedural TerminologyTable S2. Association of intraoperative fentanyl dose and postoperative respiratory complications across fentanyl-dose decilesFentanyl dose decile1st Decile(n=14,447)2nd Decile(n=14,592)3rd Decile(n=14,590)4th Decile(n=14,751)5th Decile(n=14,626)6th Decile(n=14,555)7th Decile(n=14,419)8th Decile(n=14,902)9th Decile(n=14,353)10th Decile(n=14,500)Fentanyl dose*(μg kg-1)0 (0, 0)0.95 (0.77, 1.08)1.33 (1.25, 1.41)1.65 (1.56, 1.74)2.00 (1.90, 2.08)2.37 (2.27, 2.46)2.74 (2.63, 2.81)3.13 (3.02, 3.25)3.67 (3.50, 3.85)4.72 (4.31, 5.56)Range of fentanyl dose(μg kg-1)0-00.05-1.21.2-1.51.5-1.81.8-2.22.2-2.52.5-2.92.9-3.33.3-4.04.0-76.1Postoperative Respiratory Complications Positive outcomes**1,798(12.5%)1,488(10.2%)1,389(9.5%)1,493(10.1%)1,765(12.1%)1,861(12.8%)2,094 (14.5%)2,238(15.0%)2,178(15.2%)2,535(17.5%)adjusted OR [95% CI]10.82[0.75-0.89]P<0.0010.81 [0.74-0.88]P<0.0010.86[0.80-0.94]P=0.0010.94 [0.87-1.02]P=0.1220.93[0.86-1.01]P=0.0851.05[0.97-1.13]P=0.2431.09[1.01-1.18]P=0.0291.08 [0.995-1.17]P=0.0651.15[1.07-1.25]P<0.001adjusted OR [95% CI]0.87[0.80-0.94]P<0.0010.71[0.65-0.77]P<0.0010.70[0.64-0.76]P<0.0010.75[0.69-0.81]P<0.0010.81 [0.75-0.88]P<0.0010.81 [0.75-0.87]P<0.0010.91 [0.84-0.98]P=0.0130.94 [0.88-1.02]P=0.1320.93 [0.87-1.00]P=0.0631Adjusted odds ratio (OR) and 95% confidence interval (CI) were derived from adjusted logistic regression analyses.*median (interquartile range)**Event rates of postoperative respiratory complications are shown as total numbers (percentages).Table S3. Results of exploratory outcomes across fentanyl dose quartilesFentanyl dose quartile4th Quartile(n=36,095)3rd Quartile(n=36,634)2nd Quartile(n=35,994)1st Quartile(n=37,012)Fentanyl dose (μg kg-1)3.85(3.42, 4.50)2.63(2.42, 2.88)1.74(1.54, 1.96)0.80(0, 1.14)Ileus within 3 daysPositive outcomes1,620 (4.5%)1,183 (3.2%)915 (2.5%)776 (2.1%)aOR [95% CI]10.98 [0.90-1.07]P=0.6760.95 [0.87-1.04]P=0.2480.88 [0.80-0.97]P=0.009ICU admissionPositive outcomes4,613 (12.8%)3,493 (9.5%)2,531 (7.0%)3,123 (8.4%)aOR [95% CI]10.85 [0.81-0.90]P=0.0010.75 [0.71-0.80]P=0.0010.86 [0.81-0.91]P<0.001Wound infectionPositive outcomes2,360 (6.6%)2,406 (6.6%)2,156 (6.0%)2,173 (5.9%)aOR [95% CI]10.98 [0.92-1.05]P=0.5900.96 [0.89-1.02]P=0.1820.88 [0.82-0.94]P<0.001Perioperative naloxone usePositive outcomes707 (2.0%)549 (1.5%)383 (1.1%)373 (1.0%)aOR [95% CI]10.81 [0.72-0.91]P=0.0010.65 [0.56-0.74]p P<0.0010.56 [0.48-0.64]P<0.001Total hospital costsaIRR [95% CI]10.93 [0.92-0.94]P<0.0010.90 [0.89-0.91]P<0.0010.92 [0.91-0.93]P<0.001Hospital length of stayMedian (IQR) in days4 (2, 7)3 (2, 6)2 (1, 5)2 (1, 5)aIRR [95% CI]10.92 [0.90-0.93]P<0.0010.88 [0.87-0.90]P<0.0010.92 [0.90-0.93]P<0.001Post-extubation desaturation (<90 %) (n=125,680)Positive outcomes1,400 (4.5%)1,431 (4.6%)1,395 (4.5%)1,560 (4.8%)aOR [95% CI]10.89 [0.83-0.97]P=0.0070.91 [0.84-0.99]P=0.0350.92 [0.85-1.00]P=0.059Post-extubation desaturation (<80 %) (n=125,680)Positive outcomes359 (1.2%)331 (1.1%)303 (0.98%)315 (0.98%)aOR [95% CI]10.88 [0.75-1.04]P=0.1300.88 [0.75-1.05]P=0.1490.84 [0.71-0.998]P=0.048Non-invasive ventilationPositive outcomes92 (0.25%)182 (0.50%)169 (0.47%)157 (0.42%)aOR [95% CI]11.16 [0.89-1.52]P=0.2791.11 [0.84-1.47]P=0.4511.03 [0.77-1.38]P=0.83230-day mortalityPositive outcomes216 (0.6%)171 (0.5%)202 (0.6%)291 (0.8%)aOR [95% CI]10.91 [0.74-1.13]P=0.3841.00 [0.81-1.24]P=0.9811.20 [0.98-1.47]P=0.083Constipation within 3 daysPositive outcomes446 (1.2%)339 (0.9%)312 (0.9%)316 (0.9%)aOR [95% CI]10.94 [0.81-1.10]P=0.4500.94 [0.80-1.09]P=0.3950.96 [0.82-1.12]P=0.596Delirium within 3 daysPositive outcomes646 (1.8%)478 (1.3%)387 (1.1%)437 (1.2%)aOR [95% CI]10.89 [0.78-1.01]P=0.0630.84 [0.73-0.96]P=0.0130.91 [0.79-1.04]P=0.182Event rates of postoperative respiratory complications are shown in total numbers (percentages). Adjusted odds ratio (aOR) or adjusted incident rate ratio (aIRR) and 95% confidence interval (95% CI) were estimated using adjusted logistic regression analyses, if not specified otherwise. If not further specified, the analyses were performed within the complete cohort.ICU = intensive care unit; IQR = interquartile rangeTable S4. Sensitivity analysesFentanyl dose quartile4th Quartile(n = 36,095)3rd Quartile(n = 36,634)2nd Quartile(n = 35,994)1st Quartile(n = 37,012)Considering only major respiratory complications (reintubation or respiratory failure)Positive outcomes1,705 (4.7%)1,212 (3.3%)801 (2.2%)835 (2.3%)aOR [95% CI]10.85 [0.78-0.93]P<0.0010.77 [0.70-0.85]P<0.0010.84 [0.76-0.92]P<0.001Considering only reintubationPositive outcomes174 (0.5%)131 (0.4%)96 (0.3%)105 (0.3%)aOR [95% CI]10.80 [0.64-1.01]P=0.0630.65 [0.50-0.84]P=0.0010.68 [0.53-0.87]P=0.002Excluding cases with preoperative opioid prescription (n=129,725)Positive outcomes5,282 (16.5%)4,575 (14.0%)3,527 (11.0%)3,669 (11.1%)aOR [95% CI]10.87 [0.83-0.92]P<0.0010.79 [0.74-0.83]P<0.0010.79 [0.75-0.84]P<0.001Excluding cases with intraoperative administration of long-acting opioids (n=48,769)Positive outcomes2,733 (22.4%)2,144 (17.7%)1,637 (13.4%)1,695 (13.8%)aOR [95% CI]10.92 (0.85-0.998)P=0.0460.88 (0.80-0.95)P=0.0030.87 (0.79-0.95)P=0.002Postoperative pain managementNo prescription of pain medication on discharge (n=56,675)Positive outcomes2,439 (17.3%)2,144 (15.1%)1,744 (12.3%)1,720 (12.1%)aOR [95% CI]10.91 [0.84-0.98]P=0.0110.84 [0.78-0.91]P<0.0010.82 [0.75-0.89]P<0.001Opioid prescription on discharge (n=77,106)Positive outcomes2,812 (14.6%)2,392 (12.4%)1,722 (8.9%)1,692 (8.8%)aOR [95% CI]10.89 [0.83-0.96]P=0.0020.75 [0.69-0.81]P<0.0010.74 [0.68-0.80]P<0.001Prescription of non-opioid analgesic on discharge (n=71,570)Positive outcomes2,871 (16.1%)2,370 (13.3%)1,904 (10.6%)1,860 (10.4%)aOR [95% CI]10.86 [0.80-0.93]P<0.0010.78 [0.72-0.84]P<0.0010.76 [0.70-0.82] P<0.001Fractional polynomial modellingaOR [95% CI]10.90 (0.86-0.94)P<0.0010.81 (0.76-0.85)P<0.0010.80 (0.76-0.84)P<0.001Adjusting for intraoperative use of peripheral nerve blockade, additional confounderaOR [95% CI]10.89 [0.84-0.93]P<0.0010.79 [0.75-0.83] P<0.0010.79 [0.75-0.83] P<0.001Interaction term “fentanyl dose x duration of surgery”aOR [95% CI]10.91 [0.86-0.96]P<0.0010.81 [0.77-0.86]P<0.0010.83 [0.78-0.88]P<0.001Subgroups of short and long duration of surgeryProcedural duration under 120 minutes (n=57,025)Positive outcomes1,509 (10.7%)1,168 (8.1%)984 (6.9%)1,272 (8.9%)aOR [95% CI]10.82 [0.75-0.90]P<0.0010.70 [0.64-0.78]P<0.0010.76 [0.69-0.84]P<0.001Procedural duration at least 120 minutes (n=88,710)Positive outcomes3,962 (17.9%)3,619 (16.6%)3,276 (14.8%)3,049 (13.6%)aOR [95% CI]10.93 [0.88-0.99]P=0.0220.82 [0.77-0.87]P<0.0010.80 [0.75-0.85]P<0.001Adjust for individual procedure performed – mixed effects logistic regression (n=130,420)Positive outcomes5,156 (15.7%)4,469 (13.4%)3,318 (10.3%)3,276 (10.2%)aOR [95% CI]10.90 [0.84-0.95]P<0.0010.86 [0.81-0.92]P<0.0010.88 [0.83-0.94]P<0.001Preferences of individual anaesthesiologist - mixed effects logistic regression (n=145,725)Positive outcomes5,848 (16.2%)5,058 (13.8%)3,891 (10.8%)4,042 (10.9%)aOR [95% CI]10.86 [0.82-0.91]P<0.0010.77 [0.73-0.81]P<0.0010.79 [0.75-0.83]P<0.001Adjusted for year of surgery, additional confounderaOR [95% CI]10.88 [0.84-0.92]P<0.0010.77 [0.73-0.81]P<0.0010.77 [0.73-0.81]P<0.001Excluding cases with surgery within four weeks prior to index surgery (n=136,696)Positive outcomes5,319 (15.6%)4,587 (13.4%)3,536 (10.4%)3,583 (10.4%)aOR [95% CI]10.89 [0.84-0.93]P<0.0010.78 [0.74-0.83]P<0.0010.79 [0.75-0.84]P<0.001Charlson Comorbidity Index5 (CCI)CCI < 3 (n=96,501)Positive outcomes2,202 (9.2%)1,892 (7.8%)1,344 (5.6%)1,605 (6.7%)aOR [95% CI]10.91 [0.84-0.98]P=0.0110.77 [0.71-0.83]P<0.0010.90 [0.83-0.98]P=0.011CCI ≥ 3 (n=49,234)Positive outcomes3,482 (28.5%)3,242 (26.2%)2,654 (21.5%)2,418 (19.6%)aOR [95% CI]10.91 [0.85-0.98]P=0.0070.82 [0.77-0.89]P<0.0010.74 [0.69-0.80]P<0.001ASA status classificationASA of 1-2 (n=99,591)Positive outcomes2,449 (9.8%)2,081 (8.5%)1,493 (5.9%)1,479 (5.9%)aOR [95% CI]10.87 [0.81-0.94]P<0.0010.74 [0.69-0.80]P<0.0010.78 [0.72-0.84] P<0.001ASA of 3-5 (n=46,144)Positive outcomes3,487 (30.2%)2,992 (26.0%)2,481 (21.5%)2,377 (20.6%)aOR [95% CI]10.90 [0.84-0.97]P=0.0030.84 [0.78-0.90]P<0.0010.81 [0.75-0.88]P<0.001Surgical service*High-risk surgery (n=58,298)Positive outcomes3,689 (25.4%)3,365 (23.1%)2,509 (17.2%)2,333 (16.0%)aOR [95% CI]10.93 [0.87-0.99]P=0.0430.73 [0.68-0.78]P<0.0010.75 [0.69-0.80]P<0.001Low-risk surgery (n=87,437)Positive outcomes1,971 (9.1%)1,737 (7.9%)1,549 (7.1%)1,686 (7.7%)aOR [95% CI]10.89 [0.82-0.96]P=0.0030.92 [0.85-0.99]P=0.0430.91 [0.84-0.99]P=0.032Preoperative respiratory comorbidities**Yes (n=30,681)Positive outcomes2,088 (27.3%)1,717 (22.5%)1,422 (18.4%)1,398 (18.2%)aOR [95% CI]10.90 [0.82-0.98]P=0.0180.79 [0.72-0.87]P<0.0010.79 [0.71-0.87]P<0.001No (n=115,054)Positive outcomes3,864 (13.5%)3,309 (11.6%)2.503 (8.7%)2,538 (8.8%)aOR [95% CI]10.91 [0.86-0.97]P=0.0020.81 [0.76-0.86]P<0.0010.81 [0.76-0.87]P<0.001Type of admissionAmbulatory (n=36,391)Positive outcomes326 (3.6%)303 (3.3%)270 (3.0%)334 (3.7%)aOR [95% CI]10.86 [0.72-1.02]P=0.0770.70 [0.58-0.84]P<0.0010.68 [0.56-0.82]P<0.001Inpatient (n=109,344)Positive outcomes4,998 (18.3%)4,580 (16.8%)3,993 (14.6%)4,035 (14.7%)aOR [95% CI]10.92 [0.87-0.97]P=0.0020.80 [0.76-0.85]P<0.0010.80 [0.76-0.85]P<0.001Multiple imputation (n=166,896)Positive outcomes6,591 (16.2%)5,896 (13.8%)4,507 (10.8%)4,662 (11.2%)aOR [95% CI]10.90 [0.86-0.94]P<0.0010.81 [0.77-0.85]P<0.0010.81 [0.77-0.85]P<0.001Event rates of postoperative respiratory complications are shown in total numbers (percentages). Adjusted odds ratio (aOR) and 95% confidence interval (95% CI) were estimated using adjusted logistic regression analyses, if not specified otherwise.*High-risk surgical service was defined as vascular, transplant, thoracic, general, neuro-, and burn surgery. Low-risk surgical service was defined as non-high-risk surgical service.**Preoperative respiratory comorbidities included chronic obstructive pulmonary disease, obstructive sleep apnoea, asthma, and pulmonary hypertension – defined as ICD-9/-10 diagnosis within one year prior to the index surgery.ASA status classification = American Society of Anesthesiologist physical status classification.Table S5. Interaction analysis - Body Mass IndexInteraction term “fentanyl dose (continuous) x BMI (continuous)”: p<0.001, OR 1.004, CI 1.002-1.005) Out-comeLowest fentanyl dose quartile (Q1)*Highest fentanyl dose quartile (Q4)*Estimated Probability –?Q1 (95% CI)**Estimated Probability – Q4 (95% CI)**Adjusted Absolute Risk Difference (95% CI)**Adjusted OR (95% CI)**P-value**Patients with BMI < 35 kg m-2, n=124,345PRCs3,365/31,194(10.8%)4,991/30,834(16.2%)7.0% (6.7-7.3%)8.4%(8.1-8.8%)-1.5% (-1.9 to -1.1%)0.81 (0.77-0.86)<0.001Severely obese patients (BMI ≥ 35 kg m-2), n=21,390PRCs606/5,353 (11.3%)847/5,300 (16.0%)8.0%(7.3-8.8%)9.5%(8.7-10.3%)-1.5% (-2.5 to -0.4%)0.83(0.73-0.95)0.006*Observed event rates of postoperative respiratory complications are shown in total numbers (percentages). **Estimated probability (adjusted absolute risk), adjusted absolute risk difference and adjusted odds ratio (OR) including 95% confidence intervals (95% CI) were estimated using adjusted logistic regression analyses.BMI = body mass indexTable S6. Characteristics of study population by intraoperative fentanyl dose – after propensity matching of 1st and 4th quartileCharacteristics1st Quartile(n=19,887)4th Quartile(n=19,887)Standardized mean differencesFentanyl dose (μg kg-1)0.65(0.00, 1.16)3.70(3.36, 4.36)PREOPERATIVE CHARACTERISTICSAge (years)54.3 ± 16.953.8 ± 16.9-0.028Gender (female)8,880 (44.7%)7,882 (39.6%)0.102BMI (kg m-2)27.3 ± 6.226.1 ± 5.3-0.216ASA physical status classification2.2 ± 0.62.2 ± 0.6-0.002Ambulatory surgery4247 (21.4%)3729 (18.8%)0.065Score for Prediction of Postoperative Respiratory Complications11 (SPORC) ≥ 7813 (4.1%)860 (4.3%)0.012Emergency surgery748 (3.8%)993 (5.0%)0.060High risk surgical service7904 (39.7%)8367 (42.1%)0.047Score for Prediction of Obstructive Sleep Apnoea (SPOSA)12 ≥ 253799 (19.1%)3268 (16.4%)-0.070History of arterial hypertension8097 (40.7%)7833 (39.4%)-0.027History of atrial fibrillation1668 (8.4%)1315 (6.6%)-0.067History of congestive heart failure1613 (8.1%)1550 (7.8%)-0.012History of coronary artery disease2352 (11.8%)2298 (11.6%)-0.008History of chronic pulmonary disease3318 (16.7%)3588 (18.0%)0.036History of pulmonary hypertension299 (1.5%)295 (1.5%)-0.002History of diabetes mellitus2590 (13.0%)2281 (11.5%)-0.047History of dyslipidaemia6220 (31.3%)5782 (29.1%)-0.048History of hemi- or paraplegia595 (3.0%)591 (3.0%)-0.001History of liver disease2152 (10.8%)2176 (10.9%)0.004DNR code status (Do Not Resuscitate)46 (0.23%)47 (0.24%)0.001Preoperative opioid subscription2142 (10.8%)2200 (11.1%)0.009INTRAOPERATIVE CHARACTERISTICSDuration of surgery (h)2.9 ± 1.83.3 ± 2.10.204Work relative value unit15.0(9.1, 21.9)15.8(10.5, 23.3)0.157Minimum alveolar concentration of inhalational anesthetics4 (age adjusted)0.81 ± 0.360.83 ± 0.310.047Other intraoperative opioid dose15(oral morphine equivalent, mg kg-1)*17.0 (0.0, 30.0)10.2 (0.0, 25.5)-0.048Multiples of 95% effective NMBA dose5, 132.3 (0.0, 3.6)2.7 (1.6, 4.1)0.242Neuraxial anaesthesia1028 (5.2%)1720 (8.6%)0.138Vasopressor dose14(mg norepinephrine equivalents)0.02 (0.00, 0.19)0.03 (0.00, 0.24)0.035Duration of intraoperative hypotension (minutes)0 (0, 1)0 (0, 2)0.050Intraoperative fluid infusion (ml)**1250 (900, 2000)1500 (1000, 2400)0.084Units of packed red blood cells0.065019172 (96.4%)18919 (95.1%)1331 (1.7%)406 (2.0%)2246 (1.2%)353 (1.8%)≥3138 (0.7%)209 (1.1%)Positive end expiratory pressure (cmH2O)5 (3, 5)5 (3, 5)0.054Plateau pressure (cmH2O)18.4 ± 5.018.2 ± 4.4-0.038Inspiratory O2-Fraction (%)52 (39, 60)53 (40, 60)-0.001Values are frequency (percentages), or median (interquartile range: 25th-, 75th-percentile); plus-minus values are mean ± standard deviation (SD).*Composite dose of hydromorphone, morphine, methadone, sufentanil.**Crystalloid-to-colloid ratio 1.5:1.ASA = American Society of Anesthesiologists; BMI = body mass index; NMBA = neuromuscular blocking agents.Supplemental section 7. Supplemental FiguresFigure S1a. Area under the receiver operating characteristics curve for the primary regression model independent of fentanyl doseThe area under the receiver operating characteristics (ROC) curve of the primary regression model, independent of intraoperative fentanyl dose, was 0.823 (CI 0.820-0.827), indicating sufficient model discrimination.Figure S1b. Reliability plot for the primary regression model The reliability plot indicates good calibration of the primary regression model for postoperative respiratory complications.Figure S2. Composite and individual respiratory complications by year of surgery In this figure, the rate of the composite outcome and the individual postoperative respiratory complications within 3 days after surgery by year of procedure is displayed.References ADDIN EN.REFLIST 1 Organization WH. Obesity: preventing and managing the global epidemicReport of a WHO Consultation 20002 Brueckmann B, Villa-Uribe JL, Bateman BT, et al. Development and validation of a score for prediction of postoperative respiratory complications. Anesthesiology 2013; 118: 1276-853 Shin CH, Grabitz SD, Timm FP, et al. Development and validation of a Score for Preoperative Prediction of Obstructive Sleep Apnea (SPOSA) and its perioperative outcomes. BMC Anesthesiol 2017; 17: 714 Royston P, Ambler G, Sauerbrei W. The use of fractional polynomials to model continuous risk variables in epidemiology. Int J Epidemiol 1999; 28: 964-745 Cancer pain relief: with a guide to opioid availability. Geneva, 19966 Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. Journal of Clinical Epidemiology 1994; 47: 1245-51 ................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.