Forecasting new home oxygen requirements after lung resection: A predictive model Authors: Benjamin Hambright, Wendelyn M. Oslock, Ioannis Liapis, Angel Lockridge, Sandipkumar Chaudhari, Himani Bhatt, Rongbing Xie, Lamario Williams, James Donahue, Benjamin Wei Published: 2026 ABSTRACT ObjectiveThere are currently no reliable means for accurately predicting which patients will require discharge with supplemental oxygen following lung resection. To address this problem, we endeavored to develop and validate a prediction model and risk nomogram for new home oxygen requirement after lung resection.MethodsWe performed a retrospective cohort study using an institutional Society of Thoracic Surgeons database. Adult patients who underwent lung resection for any indication from 2013 to 2023 were included. We excluded patients with preoperative home oxygen and missing pulmonary function tests. Bivariate analyses compared characteristics of patients who required home oxygen postdischarge with those who did not. Multivariable logistic regression models were used to identify risk factors for home oxygen, including the creation of a risk nomogram developed and validated via bootstrapping.ResultsOf 1875 patients, 6.4% had a new home oxygen requirement. Overall, a majority of patients (72%) were older than age 60 years, 54% were women, and 82% were White. Most patients underwent resection for lung cancer (69%). The most commonly performed procedure was a lobectomy (50%), and a minimally invasive approach was utilized in 84%. On multivariable analysis, patients with a new home oxygen requirement were more likely to be obese (adjusted odds ratio [aOR], 2.62; 95% CI, 1.67-4.11), be current smokers (aOR, 2.52; 95% CI, 1.58-3.96), have a preoperative diffusing capacity for carbon monoxide <50% of predicted (aOR, 2.88; 95% CI, 1.65-4.91), and have a preoperative forced expiratory volume in 1 second <50% of predicted (aOR, 2.14; 95% CI, 1.07-4.07).ConclusionsIn this large retrospective study, we identified 4 primary factors associated with increased odds of home oxygen requirement after lung resection: obesity, current tobacco use, reduced forced expiratory volume in 1 second, and reduced diffusion capacity of carbon monoxide. Prediction modeling may aid in preoperative risk assessment and targeting prehabilitation deployment. Risk nomogram predicting risk of home oxygen therapy based on pulmonary function tests, body mass index (BMI), and smoking. DLCO, diffusion capacity of carbin monoxide; FEV1, forced expiratory volume in 1 second. Central MessageA novel risk nomogram predicts new home oxygen needs after lung resection using preoperative DLCO, FEV1, BMI, and smoking status, aiding patient counseling, prehabilitation, and surgical planning. PerspectiveThis study addresses a critical gap in preoperative risk assessment by identifying key predictors of new home oxygen requirement after lung resection. By developing this validated nomogram for this outcome, the findings offer clinicians a practical tool to guide patient counseling, optimize prehabilitation strategies, and support informed decision making. Risk nomogram predicting risk of home oxygen therapy based on pulmonary function tests, body mass index (BMI), and smoking. DLCO, diffusion capacity of carbin monoxide; FEV1, forced expiratory volume in 1 second. Central Message A novel risk nomogram predicts new home oxygen needs after lung resection using preoperative DLCO, FEV1, BMI, and smoking status, aiding patient counseling, prehabilitation, and surgical planning. Perspective This study addresses a critical gap in preoperative risk assessment by identifying key predictors of new home oxygen requirement after lung resection. By developing this validated nomogram for this outcome, the findings offer clinicians a practical tool to guide patient counseling, optimize prehabilitation strategies, and support informed decision making. Before lung resection, a thorough preoperative workup is essential to minimizing and predicting potential postoperative respiratory complications. The cornerstone of this assessment is pulmonary function testing (PFT) via spirometry to measure forced expiration volume in 1 second (FEV1). If FEV1 is <60%, further testing is necessary, particularly calculating the predicted postoperative FEV1. If the predicted postoperative FEV1 is <40% of the predicted value, there is an increased risk of postoperative mortality.1, 2, 3 Assessing lung function via gas exchange capacity is also critical. This is accomplished by measuring the diffusion capacity of carbon monoxide (DLCO). The British Thoracic Society defines a high-risk patient as DLCO <40%.4 Patients meeting high-risk criteria should undergo cardiopulmonary exercise testing prior to surgery. Despite the critical importance of PFTs in predicting postoperative respiratory complications, the relationship between PFTs and the likelihood of being discharged with oxygen after pulmonary resection is not well understood. Furthermore, new home oxygen requirements following lung resection are not routinely reported in clinical data sets. Nonetheless, home oxygen comes with considerable financial costs and logistical challenges. A new home oxygen requirement can have a significant impact on patient's quality of life. There is currently a dearth of data on the frequency of this outcome and a lack of understanding of which factors may predict if a patient is at increased risk of having a new home oxygen requirement. Thus, the aim of this study is to further refine preoperative predictions of new home oxygen requirement following lung resection by developing a risk nomogram. Methods Study Design and Patient Population We conducted a retrospective cohort study to evaluate factors contributing to new home oxygen requirement following lung resection. Inclusion criteria included age 18 years or older and undergoing lung resection surgery for any etiology. Exclusion criteria included history of preoperative home oxygen therapy and inpatient mortality. Given the importance of PFTs in stratifying risk of lung resection generally, patients missing FEV1 and DLCO data were excluded. A combination of our institutional Society of Thoracic Surgeons database and manual chart review was employed to identify patients who underwent lung resection of any type (wedge, segmentectomy, lobectomy, or pneumonectomy) at our tertiary medical center between years 2013 and 2023. New home oxygen requirement was defined as being prescribed home oxygen therapy on initial discharge from postoperative hospitalization, with this further being stratified by continued requirement at 30 days. Measures The Society of Thoracic Surgeons database and electronic medical record data were abstracted for demographics, comorbid conditions, preoperative PFTs, operation type, surgical approach, and pathological tissue diagnosis. Intraoperative variables such as length of operation and requirement for intraoperative transfusion were also collected. Outcome data were assessed by complication rates, unexpected escalation of care (defined by unexpected intensive care unit admission during hospitalization), length of stay, requirement of chest tube on discharge, disposition, new home oxygen requirement, 30-day readmission, and 30-day mortality. The study protocol was reviewed and approved on January 5, 2024, by the institutional review board (IRB-300011884-002), and the need for informed written consent was waived by the institutional review board. Analysis Bivariate statistical analysis with Pearson χ2 and Fischer’s exact tests was employed to compare characteristics between patients that had new home oxygen requirement and those who did not. Wilcoxon rank-sum test was also employed to compare the median hospital length of stay between the 2 groups. A multivariate logistic regression analysis was then performed to identify significant preoperative factors for new home oxygen requirement. Initially, all preoperative variables that were significantly associated with home oxygen were included in the logistical regression model. A backward elimination algorithm was employed to identify the variables that most accurately predict the probability of discharge with oxygen therapy. To ensure that a potentially important variable had not been omitted during the univariable screening, the preoperative variables that were not investigated during the backward elimination step were added to the model 1 by 1. This process tested whether these variables were significantly associated with the outcome on the multivariable setting and whether they substantially influenced the coefficients of the variables of the model (Δβ >20%). The overall process resembles the purposeful selection model building strategy initially proposed by Hosmer and Lemeshow.5 The final logistic regression model was then used for the development of a risk nomogram. This nomogram was created based on the projected influence of the included variables on the need for new home oxygen requirement following lung resection. In more detail, every risk factor on the final model was assigned a numeric score (0-100) based on their odds ratio (OR) with factors with the highest influence on the outcome receiving the highest points score. By abstracting these variables, providers can calculate the risk for home oxygen needs following lung resection. The developed nomogram was validated using bootstrapping of our original dataset, 10 000 bootstrapped datasets were created via sampling with replacement and the area under the curve of the receiver operating characteristic was calculated and from this distribution a 95% CI was generated. Overall survival was assessed using Kaplan-Meier methods, with differences between groups evaluated using the log-rank test. Multivariable associations with mortality were examined using Cox proportional hazards regression, adjusting for clinically relevant demographic, comorbidity, operative, and pulmonary function variables. Hazard ratios (HRs) with 95% CIs are reported. All statistical analyses were performed using the R statistical programming language version 4.4.0 (R Foundation for Statistical Computing). Results Patient Demographic, Clinical, and Procedural Characteristics One thousand eight hundred seventy-five patients underwent lung resection during the study period, 120 (6.4%) of whom had a new home oxygen requirement. Of those discharged with new home oxygen therapy, 44 (36.7%) required this support for more than 30 days. There were no statistically significant differences in age, gender, or race between patients who had a new home oxygen requirement and those who did not (Table 1). However, clinical characteristics did differ between the 2 groups. Patients who required home oxygen were significantly more likely to be obese (46% vs 33%; P = .006) and be diagnosed with chronic obstructive pulmonary disease (53% vs 28%; P < .001), heart failure (10% vs 4.2%; P = .003), and hypertension (72% vs 62%; P = .034). Patients requiring home oxygen were also more likely to be current smokers (38% vs 23%; P < .001) and have a >40 pack-year smoking history (48% vs 31%; P = .001). In terms of PFTs, patients requiring home oxygen were less likely to have a normal preoperative FEV1 (>80% of predicted) (30% vs 56%; P < .001) or a normal preoperative DLCO (>75% of predicted) (35% vs 50%; P < .001).Table 1Preoperative characteristics for patients undergoing lung resection from 2013 to 2023Preoperative characteristicOverall(N = 1875)Oxygen therapy needed after dischargeP valueNo(n = 1755)Yes(n = 120)Age (y).33 <60523 (28)495 (28)28 (23) 60-751081 (58)1004 (57)77 (64) >75271 (14)256 (15)15 (13)Sex.29 Female1006 (54)936 (53)70 (58) Male869 (46)819 (47)50 (42)Race.34 Black304 (16)281 (16)23 (19) White1545 (82)1448 (83)97 (81) Other26 (1.4)26 (1.5)0 (0)BMI.006 <25 - normal591 (32)566 (32)25 (21) 25-30 - overweight649 (35)609 (35)40 (33) >30 - obese635 (34)580 (33)55 (46)ASA classification.12 1 and 2 - none to mild disease57 (3.0)56 (3.2)1 (0.8) 3 - severe disease1607 (86)1507 (86)100 (83) 4 - life-threatening211 (11)192 (11)19 (16)Comorbidities Heart failure86 (4.6)74 (4.2)12 (10).003 Hypertension1174 (63)1088 (62)86 (72).03 Other cardiac comorbidity584 (31)540 (31)44 (37).18 Major vascular disease242 (13)225 (13)17 (14).67 COPD548 (29)484 (28)64 (53)<.001 Interstitial fibrosis46 (2.5)40 (2.3)6 (5.0).07 End-stage renal disease8 (0.4)7 (0.4)1 (0.8).41 Diabetes386 (21)354 (20)32 (27).09 Chronic immunosuppression158 (8.4)143 (8.2)15 (13).10Prior cardiothoracic surgery519 (28)483 (28)36 (30).56Tobacco use<.001 Current smoker453 (24)407 (23)46 (38) Former smoker911 (49)859 (49)52 (43) Never smoker511 (27)489 (28)22 (18)Pack-year history.001 0 - never smokers511 (27)489 (28)22 (18) <20 - light smokers244 (13)235 (13)9 (7.5) 20-40 - moderate smokers513 (28)482 (28)31 (26) >40 - heavy smokers595 (32)537 (31)58 (48)Disease etiology.06 Benign331 (18)314 (18)17 (14)Primary lung cancer1297 (69)1203 (69)94 (78) Metastasis247 (13)238 (14)9 (7.5)FEV1<.001 >80% - normal1024 (55)988 (56)36 (30) 60-80% - mildly abnormal578 (31)529 (30)49 (41) 50-59% - moderately151 (8.1)137 (7.8)14 (12) 40-49% - severely abnormal74 (3.9)61 (3.5)13 (11) <40% - extremely severely48 (2.6)40 (2.3)8 (6.7)DLCO<.001 >75% - normal921 (49)879 (50)42 (35) 60-75% - mild reduction517 (28)489 (28)28 (23) 50-59% - moderate reduction240 (13)222 (13)18 (15) 40-49% - severe reduction128 (6.8)108 (6.2)20 (17) <40% - very severe reduction69 (3.7)57 (3.2)12 (10)Values are presented as n (%). Bolded P values indicate statistical significance (P < .05). BMI, Body mass index; ASA, American Society of Anesthesiologists; COPD, chronic obstructive pulmonary disease; DLCO, diffusion capacity of carbon monoxide. Table 1Preoperative characteristics for patients undergoing lung resection from 2013 to 2023Preoperative characteristicOverall(N = 1875)Oxygen therapy needed after dischargeP valueNo(n = 1755)Yes(n = 120)Age (y).33 <60523 (28)495 (28)28 (23) 60-751081 (58)1004 (57)77 (64) >75271 (14)256 (15)15 (13)Sex.29 Female1006 (54)936 (53)70 (58) Male869 (46)819 (47)50 (42)Race.34 Black304 (16)281 (16)23 (19) White1545 (82)1448 (83)97 (81) Other26 (1.4)26 (1.5)0 (0)BMI.006 <25 - normal591 (32)566 (32)25 (21) 25-30 - overweight649 (35)609 (35)40 (33) >30 - obese635 (34)580 (33)55 (46)ASA classification.12 1 and 2 - none to mild disease57 (3.0)56 (3.2)1 (0.8) 3 - severe disease1607 (86)1507 (86)100 (83) 4 - life-threatening211 (11)192 (11)19 (16)Comorbidities Heart failure86 (4.6)74 (4.2)12 (10).003 Hypertension1174 (63)1088 (62)86 (72).03 Other cardiac comorbidity584 (31)540 (31)44 (37).18 Major vascular disease242 (13)225 (13)17 (14).67 COPD548 (29)484 (28)64 (53)<.001 Interstitial fibrosis46 (2.5)40 (2.3)6 (5.0).07 End-stage renal disease8 (0.4)7 (0.4)1 (0.8).41 Diabetes386 (21)354 (20)32 (27).09 Chronic immunosuppression158 (8.4)143 (8.2)15 (13).10Prior cardiothoracic surgery519 (28)483 (28)36 (30).56Tobacco use<.001 Current smoker453 (24)407 (23)46 (38) Former smoker911 (49)859 (49)52 (43) Never smoker511 (27)489 (28)22 (18)Pack-year history.001 0 - never smokers511 (27)489 (28)22 (18) <20 - light smokers244 (13)235 (13)9 (7.5) 20-40 - moderate smokers513 (28)482 (28)31 (26) >40 - heavy smokers595 (32)537 (31)58 (48)Disease etiology.06 Benign331 (18)314 (18)17 (14)Primary lung cancer1297 (69)1203 (69)94 (78) Metastasis247 (13)238 (14)9 (7.5)FEV1<.001 >80% - normal1024 (55)988 (56)36 (30) 60-80% - mildly abnormal578 (31)529 (30)49 (41) 50-59% - moderately151 (8.1)137 (7.8)14 (12) 40-49% - severely abnormal74 (3.9)61 (3.5)13 (11) <40% - extremely severely48 (2.6)40 (2.3)8 (6.7)DLCO<.001 >75% - normal921 (49)879 (50)42 (35) 60-75% - mild reduction517 (28)489 (28)28 (23) 50-59% - moderate reduction240 (13)222 (13)18 (15) 40-49% - severe reduction128 (6.8)108 (6.2)20 (17) <40% - very severe reduction69 (3.7)57 (3.2)12 (10)Values are presented as n (%). Bolded P values indicate statistical significance (P < .05). BMI, Body mass index; ASA, American Society of Anesthesiologists; COPD, chronic obstructive pulmonary disease; DLCO, diffusion capacity of carbon monoxide. Preoperative characteristics for patients undergoing lung resection from 2013 to 2023 Values are presented as n (%). Bolded P values indicate statistical significance (P < .05). BMI, Body mass index; ASA, American Society of Anesthesiologists; COPD, chronic obstructive pulmonary disease; DLCO, diffusion capacity of carbon monoxide. Surgical characteristics did not differ between patients based on need for home oxygen (Table 2). In the overall cohort, lung cancer was the most common indication for surgery, accounting for 69% of cases, followed by benign disease (18%) and metastasis (13%). In terms of procedures performed, a majority of patients underwent a lobectomy (50%), followed by wedge resection (37%), segmentectomy (11%), bilobectomy (1.8%), and pneumonectomy (0.9%). Among patients undergoing lobectomy, new home oxygen use varied by resected lobe. The proportion of patients requiring home oxygen was highest after left lower lobectomy (12 out of 147 [8.2%]), followed by right lower lobectomy (13 out of 172 [7.6%]), right upper lobectomy (24 out of 333 [7.2%]), and left upper lobectomy (4 out of 185 [2.2%]; P = .068). Eighty-four percent of resections were performed minimally invasively, whereas 16% underwent an open resection. Most surgeries were completed within 1 to 3 hours (67%). Increasing operative time was strongly associated with procedure complexity (P < .001). High-complexity resections such as pneumonectomy and bilobectomy were predominantly performed in cases lasting >3 hours (81% and 76%, respectively), whereas shorter procedures were largely composed of wedge resections and segmentectomies that typically lasted 1 to 3 hours (85% and 62%, respectively).Table 2Lung resection surgical characteristics stratified by home oxygenSurgical characteristicsOverall(N = 1875)Oxygen therapy needed after dischargeP valueNo(n = 1755)Yes(n = 120)Procedure urgency.69 Elective1848 (99)1730 (99)118 (98) Urgent or emergency27 (1.4)25 (1.4)2 (1.7)Surgical approach.60 Open296 (16)275 (16)21 (18) Minimally invasive1579 (84)1480 (84)99 (83)Extent of resection.33 Pneumonectomy16 (0.9)13 (0.7)3 (2.5) Bilobectomy34 (1.8)32 (1.8)2 (1.7) Lobectomy937 (50)877 (50)60 (50) Segmentectomy199 (11)189 (11)10 (8.3) Wedge resection689 (37)644 (37)45 (38)Surgical time (h).07 <1249 (13)237 (14)12 (10) 1-31255 (67)1180 (67)75 (63) >3371 (20)338 (19)33 (28)Required intraoperative transfusion26 (1.4)22 (1.3)4 (3.3).08Values are presented as n (%). Table 2Lung resection surgical characteristics stratified by home oxygenSurgical characteristicsOverall(N = 1875)Oxygen therapy needed after dischargeP valueNo(n = 1755)Yes(n = 120)Procedure urgency.69 Elective1848 (99)1730 (99)118 (98) Urgent or emergency27 (1.4)25 (1.4)2 (1.7)Surgical approach.60 Open296 (16)275 (16)21 (18) Minimally invasive1579 (84)1480 (84)99 (83)Extent of resection.33 Pneumonectomy16 (0.9)13 (0.7)3 (2.5) Bilobectomy34 (1.8)32 (1.8)2 (1.7) Lobectomy937 (50)877 (50)60 (50) Segmentectomy199 (11)189 (11)10 (8.3) Wedge resection689 (37)644 (37)45 (38)Surgical time (h).07 <1249 (13)237 (14)12 (10) 1-31255 (67)1180 (67)75 (63) >3371 (20)338 (19)33 (28)Required intraoperative transfusion26 (1.4)22 (1.3)4 (3.3).08Values are presented as n (%). Lung resection surgical characteristics stratified by home oxygen Values are presented as n (%). Postoperative Outcomes Overall, 29% of patients had at least 1 complication and 7.1% were readmitted. Thirty-day mortality was 0.6% (Table 3). Patients who required home oxygen also experienced more unexpected escalations of care (9.2% vs 2.3%; P < .001) and had longer hospital stays (median, 3.5 vs 2 days; P < .001). Patients who required home oxygen were more likely to experience a complication (43% vs 28%; P < .001), specifically pulmonary or cardiac complications (31% vs 14%; P < .001 and 14% vs 8.1%; P = .022, respectively). This relationship was further reinforced when logistic regression controlling for the decreased FEV1 and DLCO in the home oxygen cohort continued to show a significant association between new home oxygen requirements and any postoperative complication rates (adjusted OR [aOR], 1.76; 95% CI, 1.20-2.58; P = .0038). Univariable analyses demonstrated that decreasing FEV1 was significantly associated with multiple adverse postoperative outcomes, including overall complications, pulmonary complications, prolonged hospitalization, and discharge with a chest tube (all P < .001). DLCO, although predictive of postdischarge oxygen requirement, showed a more limited association with other postoperative complications.Table 3Postoperative outcomes following lung resection stratified by home oxygenPostoperative outcomesOverall(N = 1875)Oxygen therapy needed after dischargeP-valueNo(n = 1755)Yes(n = 120)Any complication548 (29)497 (28)51 (43)<.001 Pulmonary280 (15)243 (14)37 (31)<.001 Cardiac160 (8.5)143 (8.1)17 (14).02 Neurologic17 (0.9)15 (0.9)2 (1.7).30 VTE3 (0.2)3 (0.2)0 (0)>.99 Renal207 (11)193 (11)14 (12).82 GI28 (1.5%)25 (1.4)3 (2.5).42 Infectious63 (3.4)58 (3.3)5 (4.2).60 Unexplained care escalation51 (2.7)40 (2.3)11 (9.2)<.001Hospital length of stay2.00 (1.00-4.00)2.00 (1.00-4.00)3.50 (2.00-6.00)<.001Discharge with chest tube96 (5.1)89 (5.1)7 (5.8).71Discharge location>.99 Home1846 (98)1727 (98)119 (99) Facility29 (1.5)28 (1.6)1 (0.8%)Readmitted within 30 d134 (7.1)124 (7.1)10 (8.).60Mortality within 30 d12 (0.6)10 (0.6)2 (1.7).18Values are presented as n (%) or median (interquartile range). Bolded P values indicate statistical significance (P < .05). VTE, Venous thromboembolism; GI, gastrointestinal. Table 3Postoperative outcomes following lung resection stratified by home oxygenPostoperative outcomesOverall(N = 1875)Oxygen therapy needed after dischargeP-valueNo(n = 1755)Yes(n = 120)Any complication548 (29)497 (28)51 (43)<.001 Pulmonary280 (15)243 (14)37 (31)<.001 Cardiac160 (8.5)143 (8.1)17 (14).02 Neurologic17 (0.9)15 (0.9)2 (1.7).30 VTE3 (0.2)3 (0.2)0 (0)>.99 Renal207 (11)193 (11)14 (12).82 GI28 (1.5%)25 (1.4)3 (2.5).42 Infectious63 (3.4)58 (3.3)5 (4.2).60 Unexplained care escalation51 (2.7)40 (2.3)11 (9.2)<.001Hospital length of stay2.00 (1.00-4.00)2.00 (1.00-4.00)3.50 (2.00-6.00)<.001Discharge with chest tube96 (5.1)89 (5.1)7 (5.8).71Discharge location>.99 Home1846 (98)1727 (98)119 (99) Facility29 (1.5)28 (1.6)1 (0.8%)Readmitted within 30 d134 (7.1)124 (7.1)10 (8.).60Mortality within 30 d12 (0.6)10 (0.6)2 (1.7).18Values are presented as n (%) or median (interquartile range). Bolded P values indicate statistical significance (P < .05). VTE, Venous thromboembolism; GI, gastrointestinal. Postoperative outcomes following lung resection stratified by home oxygen Values are presented as n (%) or median (interquartile range). Bolded P values indicate statistical significance (P < .05). VTE, Venous thromboembolism; GI, gastrointestinal. Overall survival (Figure 1) was not significantly different between patients requiring oxygen postdischarge and those who did not (log-rank P = .37). In multivariable Cox analysis, discharge on new supplemental home oxygen was not associated with mortality (adjusted HR [aHR], 1.07; 95% CI, 0.67-1.71). Factors independently associated with increased mortality included age older than 75 years (aHR, 1.57; 95% CI, 1.04-2.35), male sex (aHR, 1.30; 95% CI, 1.00-1.68), American Society of Anesthesiologists class IV status (aHR, 2.96; 95% CI, 1.03-8.48), major vascular disease (aHR, 1.45; 95% CI, 1.05-2.01), dialysis dependence (aHR, 6.21; 95% CI, 1.51-25.6), metastatic disease indication (aHR, 3.39; 95% CI, 1.93-5.97), and reduced DLCO (50%-59%: aHR, 1.52; 95% CI, 1.07%-2.17% and 40%-49%: aHR, 1.78; 95% CI; 1.14-2.78).Figure 1Overall survival stratified by new home oxygen requirement at discharge. Shaded areas represent 95% CI. Overall survival stratified by new home oxygen requirement at discharge. Shaded areas represent 95% CI. Predictors of New Home Oxygen Requirement and Risk Nomogram Multivariable logistic regression analysis with backward selection to identify factors associated with a new home oxygen requirement was performed using all factors found to be significant in a univariate analysis (Table 1). In addition, factors that approached statistical significance in the univariate analysis, including intraoperative transfusion, preexisting interstitial fibrosis, procedure duration, and disease etiology were also included in the model. This analysis revealed that the following 4 preoperative factors were most strongly associated with home oxygen requirement: being obese (aOR, 2.62; 95% CI, 1.67-4.11), current smoking (aOR, 2.52; 95% CI, 1.58-3.96), preoperative DLCO <50% of predicted (aOR, 2.88; 95% CI, 1.65-4.91), and preoperative FEV1 <50% of predicted (aOR, 2.14; 95% CI, 1.07-4.07) (Table 4). When patients were stratified by cumulative preoperative risk burden (DLCO <50%, FEV1 <50%, obesity, and current smoking), the observed proportion of patients discharged with new supplemental home oxygen increased in a stepwise fashion. Among patients with no risk factors, 2.6% required home oxygen at discharge, compared with 6.8% of those with 1 risk factor, 14.6% of those with 2 risk factors, and 24.2% among patients with 3 risk factors. Although few patients had all 4 risk factors (n = 5), 20% of these patients required supplemental oxygen at discharge.Table 4Significant risk factors for a new home oxygen requirementCharacteristicOdds ratio95% CIP valueDLCO <50%2.881.65-4.91<.001BMI >302.621.67-4.11<.001FEV1 <50%2.141.07-4.07.025Current smoker2.521.58-3.96<.001DLCO, Diffusion capacity of carbon monoxide; BMI, body mass index; FEV1, forced expiratory volume in 1 second. Table 4Significant risk factors for a new home oxygen requirementCharacteristicOdds ratio95% CIP valueDLCO <50%2.881.65-4.91<.001BMI >302.621.67-4.11<.001FEV1 <50%2.141.07-4.07.025Current smoker2.521.58-3.96<.001DLCO, Diffusion capacity of carbon monoxide; BMI, body mass index; FEV1, forced expiratory volume in 1 second. Significant risk factors for a new home oxygen requirement DLCO, Diffusion capacity of carbon monoxide; BMI, body mass index; FEV1, forced expiratory volume in 1 second. A predictive nomogram was then constructed using the multivariable model, with scores assigned (0-100) for each variable (Figure 2). The model discrimination was internally validated through bootstrapping. The average C-statistic produced through 10 000 resamples of the data was 0.726 (95% CI, 0.674-0.776). To use the nomogram (Figure 2), locate the patient's value for each predictor on its corresponding axis (eg, DLCO <50%: Yes/No; FEV1 <50%: Yes/No; body mass index >30: Yes/No; current smoker: Yes/No). For each predictor, draw a vertical line up to the “Points” scale at the top of the figure to determine the number of points assigned to each characteristic. Next, sum the points across all predictors to obtain a Total Points score. Finally, locate this Total Points value on the Total Points axis and draw a vertical line down to the Predicted Value scale to read the corresponding predicted probability of new home oxygen.Figure 2Risk nomogram predicting risk of home oxygen therapy based on pulmonary function tests, body mass index (BMI), and smoking. DLCO, Diffusion capacity of carbon monoxide; FEV1, forced expiratory volume in 1 second. Risk nomogram predicting risk of home oxygen therapy based on pulmonary function tests, body mass index (BMI), and smoking. DLCO, Diffusion capacity of carbon monoxide; FEV1, forced expiratory volume in 1 second. Discussion In this single-center study, we found that patients with new home oxygen requirement following lung resection were likely to experience other negative postoperative outcomes, including unexpected escalation in care, higher complication rates, and longer hospital stays. Notably, we identified several preoperative predictors for requiring home oxygen after lung surgery: reduced DLCO, lower FEV1, obesity, and current smoking status. Based on these identified risk factors, our study is the first to develop a novel risk nomogram to predict the likelihood of home oxygen requirements at discharge after lung resection. Prior studies have also found that patients with a new home oxygen requirement are also likely to experience other adverse postoperative outcomes. A single institution study at Mount Sinai reported that nearly 15% of patients were discharged with home oxygen and similarly identified obesity and DLCO as significant risk factors.6 Another study focusing on patients with lung cancer found that patients who required home oxygen were more likely to have experienced a postoperative complication.7 This association between complications and new home oxygen requirement was also found in a recent meta-analysis, which also found that prolonged length of stay was more likely to be experienced by those with a new home oxygen requirement.8 Although new home oxygen was not significantly associated with neither overall nor 30-day mortality in our cohort, the requirement for home oxygen therapy has been associated with higher mortality rates at 6 months postoperatively and a significant reduction in overall survival in 2 prior studies.9,10 In this study we identified 4 risk factors that were predictive of home oxygen requirement after lung resection: DLCO, FEV1, current smoking status, and obesity. These factors are consistent with other studies of home oxygen requirement more generally. It has been established that patients with reduced DLCO are more likely to require supplemental oxygen due to impaired gas exchange causing increased ventilatory demand.11,12 Similarly, lower FEV1 is likely to require supplemental oxygen due to airflow obstruction and air trapping.13 Current smoking status correlates with supplemental oxygen requirement because there is a higher risk for decreased oxygen saturation.14 Lastly, obesity increases the risk for supplemental oxygen requirement due to reduced lung volumes leading to mismatch in ventilation and perfusion.15,16 Additionally, decreased chest wall compliance from increased adipose tissue will increase the mechanical load and work of breathing.17 Obesity is also associated with obstructive sleep apnea, where episodes of upper airway obstruction may require supplemental oxygen.18 Lower lobe resections were more common among patients requiring new home oxygen compared with upper lobe resections, although this finding did not quite achieve statistical significance (P = .068). This trend is biologically plausible because upper lobes are more frequently influenced by smoking-related emphysema, and their resection—particularly in patients with chronic obstructive pulmonary disease—may result in less postoperative functional impairment than removal of more functional lower lobe parenchyma. Risk nomograms provide a more personalized and detailed approach to patient care, allowing for better risk stratification. This tailored assessment fosters improved goals of care discussions and informed consent. This is particularly crucial in the context of patients undergoing lung resection because the requirement for home oxygen therapy has been associated with higher mortality rates at 6 months postoperatively and a significant reduction in overall survival.9,10 By understanding and categorizing patient risk, we can more effectively identify individuals who may benefit from intensive prehabilitation that could potentially mitigate the number of patients who will require postoperative home supplemental oxygen. Current data show that patients who undergo a prehabilitation regiment, comprising moderate to high-intensity exercise, experience significant improvements in muscle strength, 6-minute walk test performance, and maximal oxygen consumption, compared with those who do not.19, 20, 21 As such, the American Association for Thoracic Surgery highlights that PFTs and exercise testing are helpful in identifying patients at higher risk for home oxygen requirement.22 Although exercise is a key component of prehabilitation, it is equally important to address modifiable behavioral risk factors, such as smoking and dietary habits. Patients who smoke at the time of surgery are at higher risk for postoperative pulmonary complications, intensive care unit admissions, and prolonged hospital stays.23 In our study, we observed that patients requiring oxygen therapy after discharge were more likely to be current smokers and experience complications. Therefore, preoperative smoking cessation is crucial, although the timing of this intervention matters. Smoking cessation >3 weeks before surgery is linked to reduced morbidity and mortality, whereas patients who quit <3 weeks before surgery face significantly higher odds of death and morbidity.24 Moreover, research shows that patients who successfully quit smoking preoperatively are more likely to maintain abstinence at the 12-month follow-up.25 These findings must be understood within the context of several limitations. Although the findings from this study are consistent with previous observations and our sample size is larger than a majority of other studies of new postoperative home oxygen,8 generalizability is limited due to the retrospective, single-center design. Additionally, mortality was too rare of an event in our cohort to be further analyzed. Another limitation of this study is that individuals without PFTs were excluded. Further, our data do not include long-term follow-up, which could add important insights relevant to postoperative recovery. Because we did not have data on the total duration of home oxygen therapy (only 30-day data), providing an estimated risk of needing permanent supplemental oxygen for patients is not possible with our nomogram. Finally, we were unable to determine the causality of the relationship of new home oxygen requirement and increased risk of complications. It is certainly possible that this relationship could be bidirectional: The incidence of a complication such as pneumonia could increase the likelihood of needing home oxygen and pulmonary vulnerability as manifested by needing home oxygen could increase the risk of complications. Conclusions Our study is the first to produce a risk nomogram for postoperative home oxygen requirements following lung resection for any reason. Patients who have worse preoperative pulmonary function test (DLCO or FEV1 <50%), body mass index >30, and currently smoke are at the highest risk of requiring home oxygen therapy. Patients with new home oxygen requirement were more likely to experience any complication, a pulmonary or cardiac complication specifically, as well as an unexpected escalation of care. The ability to estimate likelihood of new home oxygen requirement may assist with preoperative counseling of patients on expectations, identifying need for prehabilitation, and influence treatment decisions on resection versus nonsurgical therapies. Conflict of Interest Statement The authors reported no conflicts of interest. The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest.