Lung resection and new postoperative home oxygen a systematic review

Authors: Wendelyn M. Oslock, Rongbing Xie, James Donahue, Benjamin Wei

Abstract

Background

While patients are assessed for their ability to tolerate surgery through physiologic evaluations such as pulmonary function tests, ventilation-perfusion scans, and exercising testing, some patients still require home oxygen therapy after pulmonary resection. It is not well understood what the associated risk factors are, how long patients need supplemental oxygen, and if this requirement is associated with worse long-term outcomes. Given these knowledge gaps, we sought to conduct a systematic review of pulmonary resections and new postoperative home oxygen requirement.

Methods

A comprehensive search of PubMed was conducted from inception to August 2023 for studies reporting new home oxygen requirement in adults after pulmonary resection. If all participants were already on home oxygen, if no patients required home oxygen after surgery, or full text was not available in English the study was excluded. Odds ratios from available studies were statistically combined using a meta-analytic approach into a forest plot, allowing for a more precise estimate of the effect size than any single study.

Results

A total of 17 studies capturing 26,093 patients, 3,796 of whom were discharged on home oxygen, were included in the review. The majority of studies focused on patients undergoing resection for malignant etiologies. Female sex, White race, obesity, and non-adenocarcinoma histology were most commonly found to be associated with increased risk for home oxygen. Tobacco use and pulmonary function studies [e.g., forced expiratory volume in one second (FEV1) and diffusing capacity of the lungs for carbon monoxide (DLCO)] lost significance on multivariable analysis. While few of the studies captured duration and outcomes associated with home oxygen, the ones that did found that nearly 50% of patients were still on home oxygen after 6 months and that the need for home oxygen even during the first month was associated with increased risk for mortality.

Conclusions

Requiring home oxygen after surgery may be associated with worse outcomes. Additional studies on the factors associated with requiring postoperative home oxygen are needed. This would allow for more accurate preoperative counseling as well as target modifiable risk factors for optimization.

Introduction

Pulmonary resections are common thoracic procedures performed for both malignant and benign diseases. Pulmonary resections range from wedge resections or segmentectomies which remove less than a lobe to pneumonectomies that remove an entire lung. While these surgeries are most commonly performed for cancer therapy, there are benign conditions such as chronic obstructive pulmonary disease (COPD), infection, or pneumothoraces (1) that also may require resection.

Most patients undergo preoperative pulmonary function testing prior to elective lung resection. Forced expiratory volume in one second (FEV1) is typically assessed via spirometry, with a measurement of at least 1.5 L or 60% being reassuring (2-4). If preoperative FEV1 is <60%, then additional testing is often done, such as estimating predicted postoperative FEV1 (ppoFEV1) to calculate residual lung function after resection. A ppoFEV1 of at least 35–40% is typically acceptable while <30% puts patients at increased risk for requiring postoperative respiratory support and developing other respiratory complications (3). In addition to lung volumes, the gas exchange function can also be assessed by measuring the diffusing capacity of the lungs for carbon monoxide (DLCO). Predicted postoperative DLCO (ppoDLCO) levels less than 40–50% are also associated with increased risk for the patient (3). Nonetheless, with the development of minimally invasive approaches (5) and improved anesthetic techniques, surgery has become available to more patients.

The need for supplemental oxygen after surgery can extend after discharge and may be needed long term or associated with worse outcomes (6-8). Despite home oxygen therapy being a known sequela after pulmonary resection, it is unclear what the risk factors for this are, how long most patients need supplemental oxygen, and if this requirement is associated with worse long-term outcomes. Given this uncertainty, we sought to conduct a systematic review studying new postoperative home oxygen requirement after lung resection. We present this article in accordance with the PRISMA reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-24-615/rc) (9).

Methods

Search strategy

A search was performed on PubMed from inception to August 2023 for all relevant studies using the following (Thoracic OR lung OR pulmonary) AND (home oxygen) AND (resection OR pneumonectomy OR lobectomy OR segmentectomy OR wedge. All titles and abstracts were reviewed based on inclusion and exclusion criteria described below by W.M.O. Any articles that were not excluded after this review had full text assessed. The references of included articles were then reviewed to identify additional articles that met inclusion criteria (Figure 1).

Study inclusion

All study designs were included from any time period or setting if the following criteria were (I) study population*—adult human patients; (II) intervention—pulmonary resection performed for any reason (malignant or benign) regardless of extent; (III) outcomes—*new home oxygen therapy required for at least portion of study population. Studies were excluded if they were on pediatric populations or nonhuman subjects, participants did not undergo a pulmonary surgery, if all participants were already on home oxygen, or if no patients required home oxygen after surgery. Conference abstracts and studies for which full text was unavailable or not available in English were also excluded.

Data extraction

Data were compiled into a structured form that included the following basic parameters for each citation, abstract, first author last name, year published, geographic location of study, study period, sample size and portion discharged on home oxygen, reasons for surgery, study type by one reviewer. Further, details on factors associated with need for home oxygen, duration of home oxygen therapy, and outcomes for those discharged on home oxygen were also aggregated.

Data analysis

Odds ratios from available studies were statistically combined using a meta-analytic approach. This method allows for a more precise estimate of the effect size than any single study alone. For each covariate, the odds ratios and their corresponding 95% confidence intervals were determined through a weighted average, where the weight of each study is typically inversely proportional to the variance of its effect estimate. These combined odds ratios are depicted visually with a forest plot, with points indicating the central estimate and horizontal lines depicting the confidence intervals. A vertical line at odds ratio of 1 serves as a reference, indicating no effect. The plot is designed to provide a clear comparison across different covariates and studies, helping to identify which factors might influence the likelihood of patients newly requiring home oxygen after pulmonary resection. In cases where the confidence intervals were excessively wide, indicating substantial variation or uncertainty, the ranges have been truncated on the plot for clarity, with full ranges provided separately.

Results

A total of 84 potential studies were initially identified via the search strategy outlined above (Figure 1). After screening titles and abstracts, 42 studies were excluded while full texts were sought for the remaining 42 studies. After review, an additional 25 studies were excluded including a study by Sekine* et al.*, for which a full text was not available (10). A total of 17 studies were included in the review (1,2,4,6-8,11-21).

Table 1 shows the baseline characteristics of the included studies. These studies encompassed 26,193 patients, 3,807 (14.5%) were discharged on home oxygen after surgery. The majority of patients underwent resection due to a malignant 11 studies focused exclusively on malignancy, 3 include both benign and malignant etiologies, 1 focused exclusively on benign, and 2 studies did not define the etiology. Notably 6 studies focused on patients with poor pulmonary function. While the studies came from a diverse group of countries and regions, a plurality of studies came from the United States (n=7, 41.2%) with Japan being the second most common location (n=5, 29.4%). The majority of studies were retrospective cohort studies (n=11, 64.7%). All but one study shared the results of a single institution.

Factors associated with home O2

Information from the 11 studies which provided details regarding home oxygen therapy are shown in Table 2. Nine of the studies included information on factors associated with the need for home oxygen. These factors are highlighted in Table 3. The studies reported a variety of factors that may be associated with the need for home oxygen. However, there was variability across studies with many factors only studied in 1–3 articles. Only four studies included multivariate analysis examining the factors associated with need for home oxygen; these were used to create the forest plot in Figure 2 (7,12,15,17). With regard to demographic variables, age (7,15) was not significantly associated with the need for home oxygen. White race (7,15) was a significant risk factor in one multivariate analysis and female sex (7,15,17) was a significant risk factor in 2 of 3 studies. Obesity (7,15,17) was the comorbidity found to be most frequently associated with postoperative home oxygen therapy. It was found to be significant in 2 of 3 studies, while pulmonary comorbidities (7,12,15) were significant in only 1 of 3 multivariate analyses. Cardiovascular comorbidities (15,17) were not significant in two studies, and tobacco use (12,15,17) was not significant in 3 multivariate analyses but found to be significant in two univariate analyses. Diabetes (17) was only assessed by one study but was a significant risk factor in multivariate analysis.

Pulmonary function was assessed by six of the studies (2,4,11,12,15,17). Decreased preoperative FEV1 was associated with the need for home oxygen in three univariate analyses (2,4,11) but was insignificant in a fourth study (12). ppoFEV1 was only assessed in one study (17), and lower ppoFEV1 was not associated with increased odds of home oxygen need on their univariate or multivariate analysis. Decreased DLCO was associated with need for home oxygen in two univariate analyses (2,12), but lost significance within multivariate regression (12). ppoDLCO was only assessed in one study (15). Nicastri* et al. *found that patients with ppoDLCO <60% were at increased odds of requiring home oxygen on multivariate analysis. However, for models based on specific resections, only ppoDLCO <40% maintained significance for sublobar resections while ppoDLCO lost significance on multivariate analysis for lobectomies (15). Exercise inducted hypoxia (12) was only assessed by a single study but was significant on multivariate analysis for predicting postoperative home oxygen requirement. For studies that included histology within their analysis, patients with lung cancers other than adenocarcinoma were at increased risk of requiring home oxygen in 2 of 3 multivariate analyses (7,15,17).

Regarding surgical characteristics, patients who underwent sublobar resections or pneumonectomies were at higher risk for home oxygen than lobectomies (7), while thoracotomy (17) was not higher risk than video-assisted thoracoscopic surgery (VATS) on one study’s multivariate analysis. Increased blood loss (8) and lack of enhanced recovery protocols (20) were also found to be associated with a higher risk of needing postoperative home oxygen on single univariate analyses. Regarding outcomes, experiencing a pulmonary complication (7) was associated with need for home oxygen in one multivariate analysis while both longer length of stay and discharge home were associated with home oxygen need in 1 of 2 multivariate analyses (7,17).

Duration of home O2

Duration of home oxygen requirement was detailed in three of the included studies, though the studies had very different sample sizes (7,15,16). In an analysis of the Surveillance, Epidemiology, and End Results (SEER) database, Nicastri* et al. [2018] examined 3,255 patients discharged on home oxygen, 447 (13.7%) of whom no longer needed home oxygen by 1 month. Of the 3,207 patients still alive after 1 month, 1,607 (50.1%) no longer needed home oxygen by 6 months, 1,384 (43.2%) required home oxygen for more than 6 months, and 216 patients (6.7%) had died (7). In 2022, Nicastri et al. *studied 63 patients who were discharged on home oxygen but found a different distribution for weaning this support (15). By 1 month, 58.7% of patients (n=37) were no longer requiring home oxygen, while 17.5% (n=11) were weaned by 6 months and 22.2% (n=14) still required support for over 6 months.

Outcomes associated with new home O2 requirement

Three of the studies evaluated outcomes associated with new home oxygen requirements (6-8). Sekihara* et al. found that the home oxygen therapy group (n=17 of 150) had a decreased 3-year survival compared to patients not requiring home 28% compared to 63% respectively (8). The other two studies that reported outcomes, did not conduct statistical tests to comment on significance. In the study by Garzon and colleagues, only two patients were discharged on home oxygen, however both had died within 3 months (6). Nicastri et al. *found that of those discharged on home oxygen, 1.5% died within 1 month and another 6.6% (n=216) died within 6 months, 19 of whom were weaned from oxygen within a month while the other 197 were not weaned from home oxygen before their death (7).

Discussion

This systematic review has highlighted the need for further research assessing predictive factors for the need for home oxygen after surgery as well as a better understanding of how long patients may require this support and implications for prognosis. While 17 articles were identified that discussed patients requiring home oxygen after pulmonary resection, only nine studies discussed this in any detail and only four included multivariate regression analyses. Furthermore, only one of these studies had more than 500 patients included (7).

Nonetheless, a number of factors associated with need for home oxygen were found related to patient demographics, comorbidities, pulmonary function, disease characteristics, treatment characteristics, and hospital course. Patients who are White, female, obese, have diabetes or pulmonary comorbidities, or have lung cancers other than adenocarcinoma (i.e., squamous cell, large and small cell), were found to have increased odds of requiring home oxygen in two of the largest studies (7,15). Some of these findings conflicted with the findings of other studies, which may be related to differences in the sample sizes of the studies. For example, Paul and colleagues found female and obese patients did not have increased odds of home oxygen on multivariable analysis, however their study is one of the smallest with only 50 total patients and 4 whom required home oxygen (17). While many of the above-mentioned factors are nonmodifiable, obesity and diabetes, if found to remain significant after further study, are potential risk factors that may be optimized preoperatively. For the non-modifiable characteristics, some align with previously identified disparities, e.g., female sex has also been identified as an independent risk factor for lung cancer in nonsmokers and younger individuals (22).

Tobacco use was assessed within three studies (12,15,17). While it was significant in two univariate analyses, once other variables were incorporated into the model, it no longer maintained significance. This finding was similar to pulmonary function studies. Although multiple studies showed significant relationships between FEV1 and DLCO with home oxygen on univariate analyses (2,4,11,12,17), only ppoDLCO <60% was associated with increased odds of requiring home oxygen on multivariate analysis (15). Only one study assessed preoperative exercise induced hypoxia, and they found this maintained significance on multivariate analysis (12). Overall the lack of significant findings was surprising given the reliance on measures such as FEV1 and DLCO to assess patients’ ability to tolerate surgery and the association of FEV1 and DLCO with pulmonary complications in prior studies, like that done by Ceppa and colleagues (23). This may partially be explained by the small sample sizes in the majority of the studies and suggests the need for larger studies that account for pulmonary function.

With regard to surgical factors, information was not available on duration of surgery or estimated blood loss. The failure to analyze factors such as these is a significant gap in existing literature. It was notable that both sublobar resection and pneumonectomy had higher odds of requiring home oxygen compared to lobectomy (7). This may be because patients with less reserve were selected for sublobar resections preoperatively. Additionally, those who experienced pulmonary complication or had a postoperative length of stay greater than 1 week also experienced higher odds of requiring home oxygen, though it is unclear how long these patients required this additional support.

The natural history for patients requiring postoperative home oxygen is even less well-studied. Only three articles assessed the duration of home oxygen therapy and one of these was a case study (7,15,16). Nonetheless, overall 42.2% of patients were still on home oxygen 6 months after surgery, suggesting that requiring postoperative home oxygen may exert a long-term impact on quality of life (7). In addition, the three articles that evaluated long-term outcomes found that those discharged on home oxygen had increased mortality and decreased 3-year survival (6-8). Nonetheless, it is unclear what can be done after home oxygen is initiative to reduce mortality in such cases. For individuals with COPD who require supplemental oxygen, one intervention that has shown promise in reducing mortality is telemonitoring (24). This intervention could potentially be implemented for postoperative patients with new home oxygen requirement.

Conclusions

In conclusion, requiring home oxygen after surgery may be associated with worse outcomes and turn out to be a long-term reality for nearly half of those who need this additional support initially. This points to the need for additional studies on the factors associated with requiring postoperative home oxygen as this would improve our ability to predict who may need this therapy. This would allow for more accurate preoperative counseling as well as helping clinicians target modifiable risk factors for optimization prior to surgery.

Supplementary