Natural History of Lung Hernias

Lung herniation is defined as lung tissue protrusion through a chest wall defect and is an uncommon entity that may be due to multiple etiologies. It can be due to surgical, traumatic, or congenital etiologies. Chronic obstructive pulmonary disease, asthma, infection, inflammation, neoplasm, chronic steroid use, and severe coughing are examples of potential risk factors. Lung hernias have been previously reported in case reports and case series [1‒18]. However, there is a paucity of original research studies regarding lung herniation. The objective of our study is to examine lung hernias at our medical institution with particular emphasis on their natural history.

Retrospective identification of patients with lung hernias was performed by searching the terms “lung herniation,” “pulmonary herniation,” “herniation of lung,” “lung hernia” in all computed tomography (CT) clinical radiology reports from September 2010 to November 2022. We chose CT as the imaging modality in our study as it is superior to radiography for lung hernia assessment. MRI or ultrasound was not routinely performed for our patients. Several exclusion criteria were applied to the group of retrieved patients from the initial search query, as shown in Figure 1. Since a primary study objective was to evaluate the natural history of lung herniation over time, patients were excluded if they did not undergo a follow-up CT exam or if they already underwent repair of their lung hernia prior to the follow-up CT. Patients were also excluded if they did not have lung herniation when their imaging was directly reviewed. The initial CT chest closest in time to the patient’s onset of visible hernia was defined as baseline CT. The follow-up CT scan was defined as the CT scan performed after the baseline CT. If there was more than one follow-up CT for a patient, then the CT which had the greatest duration of time since the baseline CT was designated as the follow-up CT. Baseline scans were performed on Siemens Definition (n = 23), GE Lightspeed (n = 11), Siemens Sensation (n = 2), Philips Brilliance (n = 1), and GE Revolution (n = 1) CT scanners. Follow-up scans were performed on Siemens Definition (n = 22), GE Lightspeed (n = 10), Canon Aquilion Precision (n = 4), Siemens Sensation (n = 1), and GE Revolution (n = 1) CT scanners. Both baseline and follow-up CT scan acquisitions were performed at end inspiration with the exception of two baseline scans acquired at expiratory phase. Thirty-one out of 38 baseline CT scans were contrast-enhanced exams, and 7 out of 38 baseline CT scans were noncontrast exams. Twenty-six out of 38 follow-up CT scans were contrast-enhanced exams, and 12 out of 38 follow-up CT scans were noncontrast exams. The scan range included the thorax for both baseline and follow-up CT scans except for 1 patient who did not have a follow-up CT chest but instead a CT abdomen and pelvis which included the lung bases and the lung herniation in the field of view. Baseline CT scan slice thickness ranged from 0.63 to 5 mm; however, the majority of baseline CT scans (35/38) ranged from 1 to 1.5 mm; the remaining three scans had a slice thickness of 0.63 mm, 2 mm, and 5 mm. Follow-up CT slice thickness ranged from 1.0 to 1.5 mm.

Lung hernia defect size was measured on axial CT images. Maximal anteroposterior and transverse diameters of the herniated lung parenchyma beyond the chest wall were also measured on axial CT images. Maximal craniocaudal diameter of the herniated lung parenchyma beyond the chest wall was measured on coronal CT images. An example of such measurements is shown in Figure 2.

Paraspinal muscle measurements of size and attenuation were used to derive biomarkers of muscle mass and strength. Right and left paraspinal muscles at the level of T12 were manually segmented. Segmentation was performed on CTs without contrast only as contrast would alter the actual muscle measurements. If a patient’s baseline or follow-up CT was contrast enhanced but also underwent additional noncontrast CT, then the noncontrast CT was utilized for segmentation. Data recorded included the mean muscle attenuation in HU and the standard deviation (SD) around the delineated voxels, in addition to the surface area measured in centimeter square (cm²). The mean attenuation and SD were averaged between the right and left, while the surface area was summed and corrected to the square of height in meters to get a T12 version of skeletal muscle index (SMI) measured in cm²/m². The electronic medical record was utilized to extract information regarding patients, histories, and clinical symptoms.

The primary outcome measure was hernia size progression on the last follow-up CT, defined as a 10% increase in the sum of all three diameters. Regression was defined as >10% reduction of the sum of the 3 diameters. Stable hernia has ±10% change in the diameters.

Continuous data were summarized as mean ± SD. Qualitative characteristics were summarized as frequencies and percentages. Continuous data were skewed; accordingly, Kruskal-Wallis or Mann-Whitney U tests were used to compare continuous variables between three or two categories. Analysis of the receiver operating characteristics (ROC) was used to define the best cut-off point for continuous variables that maximizes the accuracy of predicting hernia size progression.

Qualitative data were compared among groups using χ² test or Fisher’s exact test as appropriate. Univariable analyses employing Cox proportional hazard model were used to study time to progression. The time was defined as the interval between the baseline CT and the last follow-up CT. The event was defined as hernia size progression, as previously defined. Stable or regressed hernias at follow-up were included in one category.

Correlation matrix of the relevant continuous and ordinal variables was established using Spearman correlation. In all analyses, a two-tailed p value of <0.05 was considered significant.

A total of 38 patients (age: 63 ± 8.9 years, range: 34–84; 14 female) were deemed eligible for this analysis. Most of the patients were overweight or obese (82%), smokers (82%), and had a history of thoracic surgery (79%). A minority of patients had a history of trauma or forceful coughing (2/38 and 2/38, respectively). Twenty-two of 38 patients had underlying inflammatory or neoplastic conditions. A history of COPD or asthma was demonstrated in 55% and 21%, respectively (Table 1). Fourteen of 38 patients had COPD and/or asthma with a neoplastic or inflammatory condition. The 2 patients with severe or forceful coughing had either COPD or asthma. Congenital etiology or long-term corticosteroid therapy were not found in any patient.

Patients were mostly asymptomatic (33/38; 87%). The remaining 5 patients presented with chest wall pain. One of the 5 patients with chest pain had imaging evidence of acute strangulation on baseline and follow-up CT scans which were performed with contrast and showed convergence and engorgement of pulmonary vessels. The remaining 4 of 5 patients with chest pain had no imaging evidence of acute strangulation, three of whom underwent baseline and follow-up CT scans with contrast and one of which underwent noncontrast baseline and follow-up CT scans. Nineteen of 38 patients had herniation involving the anterior intercostal chest wall, and lung hernias in our study were slightly more frequently located on the left side (55%). Further details are summarized in Table 2.

The interval between the two CTs varied from 2 days to 16.8 years (median follow-up of 3.8 years). Only six hernias regressed by >10%, 20 were stable, and 12 showed size progression.

Among the clinical features, only age was significantly associated with the course of hernia (Table 3). ROC analysis identified the age of 60-year-old as the best discriminative cut-off point predictive of 10% progression of hernia diameters. Nine out of 14 patients at or under the age of 60 had hernia size progression on follow-up, while most patients above 60 (21 out of 24; 88%) had predominantly stable (n = 20) or spontaneously regressed (n = 1) hernias (p = 0.003). This association was also persistent in time-to-progression analysis. Younger age (<60 years) was marginally predictive of shorter time to progression (HR = 0.27, 95% CI: 0.07–1.04; p = 0.057). In multivariable model that included age, sex, race, changes in BMI, and the presence of inflammation or neoplasm, only age as a continuous parameter was inversely predictive of 10% progression of hernia diameters (odds ratio = 0.970; 95% CI: 0.949–0.992; p = 0.017). Within the 30 patients who had thoracic surgery, eighteen had underlying asthma or COPD; 7 of them (23%) had hernia progression, compared to 3 patients (10%) who had surgery without underlying asthma or COPD (25%). The difference, though possibly clinically meaningful, was not statistically significant (p = 0.7). It is worth noting that change in T12 paraspinal SMI was marginally higher in patients whose hernia size progressed (average increase of 4 ± 13% vs. average decrease of −6.8 ± 16.7%, p = 0.07). Also, on average, patients with progressed hernias demonstrated deterioration of muscle attenuation (−9 ± 30%, compared to an increase in muscle attenuation of 46 ± 137%). Together, the findings may suggest fatty infiltration of the skeletal muscle during the follow-up duration.

Figure 3 shows a correlogram generated from different hernia and patients’ characteristics. Men had significantly larger defects and hernia dimensions at baseline and follow-up (Table 4). However, there was no difference between men and women regarding the percentage changes for either the defect or hernia itself.

On the other hand, women tended to have more additional hernias on baseline CT (5 out of 9 compared to 2 out of 24; p = 0.077). Furthermore, inflammatory or neoplastic chest lesions were significantly more frequent in women (12 out of 14 compared to 10/24 in men; p = 0.016). Women, on average, had lower muscle attenuation and tended to have higher SD on follow-up CT; however, there were no differences between men and women at baseline, and the percentage change in the muscle metrics did not vary significantly (Table 4).

Percentage change in T12 muscle attenuation demonstrated moderate positive correlation with baseline hernia defect size (rho = 0.456) and summed hernia dimensions (rho = 0.345). Follow-up muscle attenuation demonstrated moderate negative correlation with changes in hernia dimensions (rho = −0.314). However, all these associations were not statistically significant (p values of 0.066, 0.18, and 0.166, respectively).

In this study, we have investigated lung herniation and the natural course of this condition. Literature regarding lung herniation has been comprised of case reports, case series, or reviews of reported cases. A relatively large study of lung hernias at a medical institution was published by Athanassiadi et al. [19] and consisted of 16 patients [19]. Our study, however, consisted of 38 patients and is the largest study at a single medical institution of lung hernias to the best of our knowledge. Furthermore, we have used computed tomographic imaging as both the baseline and follow-up method to assess the natural history of lung hernias. We are not aware of a systematic study in the existing literature using this method to follow up on lung hernias.

Most of our patients with lung hernias were obese or overweight and had a history of tobacco use which we suspect are likely risk factors. Additionally, prior surgery to the thorax in many patients could have led to intercostal thoracic wall musculature weakness and contributed to the development of lung hernia. Most patients in our study did not report symptoms associated with their lung herniation; therefore, we deduce the clinically asymptomatic nature of lung hernias in most cases.

The natural course of lung hernias in our study tended to mostly be either stability or size progression over time. Of note, age was predictive of hernia size progression; specifically, 60 years was the best cut-off predictive of size progression. Several characteristics demonstrated potential association with age: for example, patients with inflammation/neoplasm (60 ± 9 vs. 66 ± 8 years; p = 0.04) or asthma (57 ± 11 vs. 65 ± 8 years; p = 0.09) were younger. Also, males were slightly younger compared to female patients (61 ± 8 vs. 66 ± 10 years; p = 0.2). Men in this cohort were having significantly larger hernia defect size. We hypothesize that several of these features, in addition to some uncaptured features (e.g., level of physical activity), might contribute to the higher rate of hernia size progression seen in younger age. Separating the interaction of all the variables, via, for example, mediation analysis, was not possible in this cohort due to sample size limitations. However, the findings from our study could inform larger future studies.

Surface area and attenuation of skeletal muscles have been used as surrogate markers for sarcopenia and were deemed independent predictors of unfavorable outcomes in different diseases. Our findings support that lung hernia progression might be associated with worse muscle status. Fatty infiltration of skeletal muscles could increase the unthresholded muscle surface area (reflected in a higher SMI) and decrease the muscle attenuation due to the known lower attenuation properties of adipose tissue. We demonstrated that the continuous decline in T12 muscle attenuation (measured as percentage change in attenuation), correlates positively with baseline hernia defect size (rho = 0.456) and summed hernia dimensions (rho = 0.345). This may reflect that the insult that caused muscle status deterioration could be contributing to both initiation and progression of hernia. However, further validation of this hypothesis is warranted.

Our study is limited by our sample size as well as its retrospective design. It is also limited by the wide range of follow-up times, with sample size likely contributing to this variability. Future investigations with a greater number of patients and involving multiple institutions may yield more generalizable data.

Pulmonary herniation is an abnormal, unusual condition which can have several risk factors. Most lung hernias either remain stable or show size progression over time. Younger age was predictive of hernia size progression, with 60 years of age being the cut-off. This study may provide information to patients with this condition as well as inform clinicians, and it may have potential management implications.

The University of California Davis Institutional Review Board approved the retrospective study protocol (IRB ID-1868484-1; approval date April 8, 2022) and waived informed consent.

Lorenzo Nardo is the principal investigator (PI) of a service agreement with United Imaging Healthcare, site PI of clinical trials supported by Novartis Pharmaceuticals Corporation, PI of a clinical trial supported by Telix Pharmaceuticals, PI of a clinical trial supported by Lantheus Medical Imaging, and PI of a clinical trial supported by GE Healthcare. None of the other authors have any conflicts of interests.

The work described in the manuscript did not receive funding from any funding agency.

Mohammad H. Madani conceived the study, collected the data, and contributed to the writing of the manuscript. Yasser G. Abdelhafez performed the statistical analysis and contributed to the writing of the manuscript. Lorenzo Nardo supervised the study, reviewed the manuscript, and provided editing of the manuscript.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.