Author + information
- Received April 13, 2011
- Revision received June 17, 2011
- Accepted July 5, 2011
- Published online October 4, 2011.
- Christopher Naoum, MBBS⁎,
- Gregory L. Falk, MBBS†,
- Austin C.C. Ng, MBBS, MMed⁎,
- Tony Lu, MBBS‡,
- Lloyd Ridley, MBBS‡,
- Alvin J. Ing, MD§,
- Leonard Kritharides, MBBS, PhD⁎ and
- John Yiannikas, MBBS⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. John Yiannikas, Department of Cardiology, Concord Repatriation General Hospital, Hospital Road, Concord, New South Wales, 2139, Australia
Objectives The purpose of this study was to determine the association between cardiac compression and exercise impairment in patients with a large hiatal hernia (HH).
Background Dyspnea and exercise impairment are common symptoms of a large HH with unknown pathophysiology. Studies evaluating the contribution of cardiac compression to the pathogenesis of these symptoms have not been performed.
Methods We collected clinical data from a consecutive series of 30 patients prospectively evaluated with resting and stress echocardiography, cardiac computed tomography, and respiratory function testing before and after laparoscopic HH repair. Left atrial (LA), inferior pulmonary vein, and coronary sinus compression was analyzed in relation to exercise capacity (metabolic equivalents [METs] achieved on Bruce treadmill protocol).
Results Exertional dyspnea was present in 25 of 30 patients (83%) despite normal mean baseline respiratory function. Moderate to severe LA compression was qualitatively present in 23 of 30 patients (77%) on computed tomography. Right and left inferior pulmonary vein and coronary sinus compression was present in 11 of 30 (37%), 12 of 30 (40%), and 26 of 30 (87%) patients, respectively. Post-operatively, New York Heart Association functional class and exercise capacity improved significantly (number of patients in New York Heart Association functional classes I, II, III, and IV: 6, 11, 11, and 2 vs. 26, 4, 0, and 0, respectively, p < 0.001; METs [percentage predicted]: 75 ± 24% vs. 112 ± 23%, p < 0.001) and resolution of cardiac compression was observed. Absolute change in LA diameter on the echocardiogram was the only independent cardiorespiratory predictor of exercise capacity improvement post-operatively (p = 0.006).
Conclusions We demonstrate, for the first time, marked exercise impairment and cardiac compression in patients with a large HH and normal respiratory function. After HH repair, exercise capacity improves significantly and correlates with resolution of LA compression.
Large hiatal hernias (HHs) are associated with significant morbidity and mortality (1,2). Laparoscopic repair is generally recommended in symptomatic patients due to the risk of life-threatening complications from mechanical obstruction and incarceration (3). Symptoms typically reported include postprandial pain, heartburn, dysphagia, vomiting, and anemia (4). Dyspnea, although common, is an often misdiagnosed or unrecognized symptom in these patients (5). Occasionally, dyspnea is incorrectly attributed to comorbid respiratory or cardiac insufficiency, and patients are denied surgery on the basis of a perceived higher operative risk (6).
Despite the high prevalence of dyspnea in patients with a large HH, its mechanism is unclear. HHs can enlarge with time and cause significant compression and displacement of adjacent intrathoracic structures. Although the effects on respiratory function, diaphragmatic motility, ventilation, and perfusion are intuitively apparent, previous reports quantifying these effects are limited (6–10). Gastroesophageal reflux, which is common in HH and associated with respiratory symptoms, appears to have no effect on exercise capacity (11).
The left atrium and pulmonary veins are anatomically adjacent to the enlarging HH (Fig. 1). Case reports and 2 small case series have described cardiac compression by HH, leading to acute heart failure and hemodynamic compromise (12–17). However, there are no systematic studies evaluating the role of cardiac compression in HH-associated dyspnea.
The aim of the present study was to evaluate the functional significance of cardiac compression in relation to dyspnea and exercise capacity and in relation to the recovery of these after laparoscopic HH repair.
We prospectively collected clinical data from a consecutive series of patients with a large HH undergoing surgical repair to treat gastrointestinal symptoms (Table 1) (18–22) and prevent mechanical complications. Endoscopic or barium swallow evaluation identified patients with a large HH, defined as a hernia consisting of ≥30% intrathoracic stomach (3). The decision to undertake surgery was made before cardiorespiratory assessment, which included resting and exercise echocardiography, cardiac computed tomography (CT), and respiratory function testing before and after surgery, performed as part of the clinical evaluation. Demographic details, New York Heart Association (NYHA) functional class, symptoms, and comorbidities were collected. Clinical assessment, imaging analyses, and respiratory function testing were independently performed by separate observers. Stress echocardiography was, by clinical necessity, performed without blinding to the resting echocardiographic data. Post-operative imaging was performed knowing that patients had undergone surgery because HHs were immediately apparent (pre-operatively) or absent (post-operatively) on the echocardiogram and CT. Laparoscopic repair involved total sac excision, suture repair of the crural pillars, esophagogastropexy, and 360° fundoplication. Changes in cardiac and respiratory parameters after surgery and their relationship to exercise capacity were evaluated. The institutional Human Research Ethics Committee approved analysis and reporting of data, for which patients provided written informed consent.
Cardiac CT assessment
Contrast-enhanced cardiac CT was performed using a dual-source, 64-slice helical CT scanner (Somatom Definition, Siemens AG, Erlangen, Germany). One liter of water was orally ingested to distend the HH before image acquisition. The scan was performed during a single breath hold at end-inspiration and after injection of 100 ml of intravenous contrast media (Ultravist 370, Bayer, Leverkusen, Germany). Gating to the cardiac cycle was retrospective with images reconstructed and transferred to a workstation (Syngo 3D, Siemens Medical Solutions, Forchheim, Germany) for offline analysis, using the end-systolic phase.
Compression of the left atrium was qualitatively graded (none, mild, moderate, severe) based on the appearance of the left atrial (LA) contour immediately adjacent to the HH in 2 orthogonal views (Online Fig. 1). A convex appearance in both views was considered normal; flattening in 1 view was considered mild compression; flattening in both views was considered moderate compression; and a concave appearance in either view was considered severe compression. LA volume was quantified using 3-dimensional semiautomated software (Syngo Volume, Siemens Medical Solutions). Boundaries of the left atrium were manually traced in multiple axial slices, with exclusion of the pulmonary veins and LA appendage. The software then automatically interpolated regions between the defined regions of interest and calculated volume.
Pulmonary vein compression was qualitatively present if the caliber of the inferior pulmonary vein (IPV) adjacent to the HH appeared significantly reduced compared with the caliber of the superior pulmonary vein on the same side based on previous studies suggesting similar dimensions of the superior pulmonary vein and IPV (23). Minimal and maximal IPV diameters were measured between the ostium and first branch point in 2 orthogonal views to quantify IPV compression, with reporting of the average of these dimensions for each side. Coronary sinus (CS) compression was subjectively assessed (Figs. 2A and 2B) as reported previously (24).
HH volume was quantified using the semiautomated method described previously for LA volume calculation. Left ventricular volumes were calculated (Syngo Circulation, Siemens Medical Solutions) using threshold-based, region-growing, 3-dimensional segmentation of the left ventricle.
Doppler echocardiography assessment
Transthoracic echocardiography was performed (Vivid 7, GE Healthcare, Horten, Norway) with patients instructed to consume a meal 1 h before undergoing echocardiography. Images were stored digitally for offline analysis (Echo Pac PC, GE Healthcare). The left ventricular ejection fraction was calculated using Simpson's biplane method (25). The right ventricular systolic pressure was derived from the peak velocity of the tricuspid regurgitant envelope obtained from continuous Doppler imaging (26).
LA compression severity was qualitatively determined (none, mild, moderate, severe) by visual assessment of 4 standard echocardiographic views (Online Fig. 2). This was quantified by measuring the anteroposterior LA diameter at the LA midpoint in the parasternal long-axis view 1 frame before opening of the mitral valve (Figs. 2C and 2D). Peak systolic and diastolic pulse-Doppler velocities at the LA inflow were recorded (Figs. 2E to 2H).
Exercise testing was performed according to the maximal Bruce treadmill protocol with metabolic equivalents (METs) achieved calculated based on the duration of exercise and exercise capacity defined as the METs achieved expressed as a percentage of age-predicted values (27,28). Stress echocardiographic images were obtained immediately after exercise for offline analysis of regional wall motion to exclude myocardial ischemia. The change in exercise capacity after HH repair was measured as the difference between the post-operative exercise capacity and the pre-operative exercise capacity.
Respiratory function testing
Spirometry, lung volumes (via body plethysmography), and single-breath diffusing capacity of lung for carbon monoxide were measured according to American Thoracic Society criteria using a commercially available system (Vmax Encore, SensorMedics, Yorba Linda, California) (29–31). Respiratory function testing was often performed before the CT scan on the same day, and was always performed without oral loading.
Continuous data are expressed as mean ± SD for normally distributed variables and median (interquartile range) otherwise. Parametric and nonparametric analyses (paired and Wilcoxon matched t tests, chi-square test for trend) were performed to compare variables before and after surgery, as appropriate. Our primary hypothesis was that LA compression severity correlated with the degree of exercise capacity improvement. We therefore pre-specified LA compression severity into 4 groups (none, mild, moderate, severe). One-way analysis of variance for linear trend was used to compare means among multiple groups. A 2-tailed p < 0.05 was considered significant.
The improvement in exercise capacity after HH repair was evaluated in relation to the improvement in cardiac and respiratory variables after HH repair, using linear regression analysis. Multiple linear regression analysis, including parameters that improved significantly post-operatively and had a univariate p < 0.05, was performed to identify changes in cardiorespiratory variables that independently predicted exercise capacity improvement. Within each category (cardiac and respiratory), less significant univariables correlating significantly (R > 0.6) with other univariables in the model were removed to avoid multicollinearity. Statistical analysis was performed using Prism 5.0b and Instat 3.1a (GraphPad Software, La Jolla, California).
Between 2009 and 2010, 52 consecutive patients with large and symptomatic HHs underwent laparoscopic repair, of whom 30 underwent all pre-operative and post-operative investigations and were included in the final analysis (Fig. 3).
Baseline characteristics are summarized in Table 1. Exertional dyspnea was reported by the majority of patients (25 of 30, 83%) despite minimal rates of significant cardiac or respiratory disease. The mean left ventricular ejection fraction was normal (61 ± 9%), and there were no patients with a left ventricular ejection fraction <40% or inducible myocardial ischemia. Two patients had moderate valvular heart disease. Mean spirometric, lung volume, and diffusing capacity of lung for carbon monoxide values, as percentages of predicted, were normal at baseline. A history of anemia was reported in 9 of 30 patients (30%); however, the measured hemoglobin at pre-operative assessment was normal for the overall cohort (136 g/l [interquartile range: 131 to 148 g/l).
Laparoscopic repair was achieved in all but 2 patients who required conversion to open laparotomy. Complications occurred in 3 patients. A wound infection and post-operative respiratory failure requiring invasive ventilation developed in 1 patient. One sustained inadvertent colonic injury during dissection of adhesions. The third patient had a perioperative myocardial infarction, despite normal findings on the pre-operative stress echocardiogram. Coronary angiography was subsequently performed, which demonstrated minor coronary artery disease. There were no deaths in-hospital or at 30 days.
Cardiac CT and Doppler echocardiography findings
CS compression was the most frequently detected abnormality (26 of 30, 87%). Moderate to severe LA compression was observed in the majority of patients by qualitative assessment on cardiac CT (none: 4 of 30, 13%; mild: 3 of 30, 10%; moderate: 14 of 30, 47%; severe: 9 of 30, 30%) and echocardiogram (none: 2 of 30, 7%; mild: 8 of 30, 27%; moderate: 13 of 30, 43%; severe: 7 of 30, 23%). The right and left IPVs were compressed in 11 of 30 patients (37%) and 12 of 30 patients (40%), respectively. There were high levels of interobserver agreement for qualitative assessments (Online Figs. 1 and 2).
Quantitative cardiac parameters are presented in Table 2. On cardiac CT, LA volume and left and right IPV diameters improved significantly after surgery. In patients with severe LA compression at baseline, there was a significant increase in left ventricular end-diastolic and end-systolic volumes after surgery, consistent with improved ventricular filling. On the echocardiogram, the LA diameter increased significantly after HH repair (25 ± 8 mm vs. 36 ± 6 mm, p < 0.001). The pre-operative atrial inflow peak systolic and diastolic velocities were 88 ± 24 cm/s and 72 ± 23 cm/s, respectively. These decreased significantly after surgery to 74 ± 16 cm/s (p = 0.003) and 54 ± 16 cm/s (p = 0.004), respectively. The absolute reduction in systolic atrial inflow velocity after surgery was greater in patients with moderate to severe versus none to mild LA compression (21 ± 23 cm/s vs. 1 ± 22 cm/s, p = 0.03).
Airflow limitation was present in 9 of 30 patients (30%); however, this was only mild in 7 of 30 patients (23%) and moderate in 2 of 30 patients (7%). Abnormal lung volumes were present in 4 of 30 patients (13%) (Table 1). Post-operatively, there was a modest but statistically significant increase in forced expiratory volume in the first second of expiration (FEV1), forced vital capacity (FVC), and total lung capacity (Table 3).
Exercise capacity and functional class
Exercise capacity was reduced at baseline to 75 ± 24% of predicted (6.2 ± 2.4 METs), increasing to 112 ± 23% of predicted (9.2 ± 2.4 METs) after HH repair (p < 0.001). NYHA functional class was similarly impaired pre-operatively and improved after surgery (Fig. 4).
The change in exercise capacity after surgery was greater in patients with a higher grade of qualitative LA compression at baseline on either echocardiography or cardiac CT and in patients with right IPV compression. A similar trend was observed in patients with compression of the left IPV; however, this was not statistically significant. CS compression did not predict a greater change in exercise capacity (Fig. 5).
Linear regression analysis demonstrated a significant correlation between the improvement in exercise capacity after surgery and changes to the LA diameter on echocardiography; right IPV diameter on CT; and FEV1 (percentage predicted), FVC (percentage predicted) and diffusing capacity of lung for carbon monoxide (percentage predicted) on respiratory function tests. Multivariable linear regression analysis identified the only independent predictor to be change in the LA diameter on echocardiogram (p = 0.006) (Table 4).
The present study identifies for the first time that significant exercise impairment and extrinsic cardiac compression are common in patients with a large HH despite normal mean baseline respiratory function. The recovery of exercise capacity and cardiac parameters after laparoscopic repair indicate that large HHs directly impair exercise capacity that is due, in part at least, to cardiac compression.
A major finding of the present study is that dyspnea is common among patients with a large HH presenting for surgical repair, with 83% of patients reporting this symptom in our group. Although Low and Simchuk (6) reported a similar prevalence of 84% in their cohort, most large studies of HHs do not describe dyspnea and exercise impairment as prominent symptoms (4,5). This is important because laparoscopic surgery is generally reserved for symptomatic patients (3). Importantly, our patients had minimal evidence of severe cardiac or respiratory disease at baseline, indicating that exercise tolerance and dyspnea cannot be simply attributed to comorbidities. Moreover, patients represented their own controls, such that any observed change in exercise capacity after surgery can be confidently attributed to surgical correction of the HH.
Previous studies suggest that dyspnea in HHs is predominantly due to a mechanical respiratory effect of a large space-occupying intrathoracic mass. In the study by Low and Simchuk (6), only mild abnormalities of spirometry were identified (FEV1 [percentage predicted] and FVC [percentage predicted], 76% and 79%, respectively), despite moderately severe symptoms. After surgery, dyspneic symptoms completely resolved in most of their patients; however, there was only a mild improvement in spirometric values (absolute increase in both FEV1 [percentage predicted] and FVC [percentage predicted] of 13%) (6). In the present study, we demonstrated normal mean spirometry at baseline, despite significant dyspnea and exercise impairment. Studies correlating respiratory symptoms with spirometry suggest that the likelihood of reporting dyspnea only increases significantly in the general population when FEV1 (percentage of predicted) decreases to <75%, which is significantly lower than the mean baseline value in our study (32). After surgery, there was a significant improvement in spirometry (9% and 13% absolute increase in FEV1 [% predicted] and FVC [% predicted], respectively); however, such degrees of improvement, when observed in patients receiving bronchodilator therapy for chronic obstructive pulmonary disease, do not result in the magnitude of exercise capacity improvement observed in our cohort (33).
Unlike the relatively mild impairment of basal respiratory function in some of our patients, most demonstrated moderate to severe LA compression on echocardiography and cardiac CT, both qualitatively and quantitatively. Many patients also demonstrated compression of the IPVs and increased pulse-Doppler velocity at the LA inflow, implying that extrinsic compression was functionally important. After surgery, resolution of these cardiac abnormalities was accompanied by significant improvement in dyspnea and exercise capacity. The degree of improvement in exercise capacity with surgery correlated with the qualitative assessment of pre-operative LA compression severity (Fig. 5).
The relationship among the recovery of exercise capacity, the resolution of cardiac compression, and improvement in respiratory function were explored in this study. Multivariable regression analysis demonstrated that the only independent predictor of improved exercise capacity after HH repair was the change in the LA diameter on echocardiography (Table 4). Changes in respiratory variables after HH repair did not independently predict exercise capacity improvement. These data, coupled with the normal mean baseline respiratory function, suggest that although respiratory dysfunction contributes to HH-associated dyspnea, extrinsic cardiac compression may play a more important role in the pathogenesis.
LA compression may cause dyspnea by increasing the pulmonary venous pressure, producing interstitial edema and reduced pulmonary compliance. Previous case reports describing cardiac failure and dyspnea attributable to LA compression by HH support this hypothesis (16,34). Echocardiographic demonstration of pulmonary vein compression by various pathologies has been described in previous case reports, but not in patients with large HHs (35,36). These reports demonstrated increased systolic and diastolic components of the pulse-wave Doppler signal at the pulmonary vein ostium. Our findings of increased velocities at the LA inflow, which resolve significantly after surgery, may represent a similar pathophysiologic process related to compression by the HH.
Extrinsic cardiac compression also appears to have an effect on left ventricular filling because patients with severe LA compression had improved ventricular volumes after HH repair. Case reports of HH causing hemodynamic instability including hypotension requiring inotropic therapy or resulting in syncope are consistent with these findings (14,37–39). Impaired ventricular filling due to LA compression may also contribute to exercise intolerance by preventing the necessary increase in cardiac output that normally occurs with exercise.
Compression of the CS was present in 87% of patients. The anatomic course of the CS in the posterior atrioventricular groove makes it particularly susceptible to compression. However, CS compression was not a useful predictor of improved exercise capacity after surgery. This may be due to the observation that with a binary qualitative method for determining CS compression, most patients demonstrate this abnormality, making it a nondiscriminatory variable in this population. CS compression leading to diastolic dysfunction and dyspnea has been described in a patient with lymphoma, but not in patients with a HH (40). Previous animal studies have confirmed a relationship among CS compression and impaired myocardial blood flow, increased ventricular blood volume, decreased ventricular distensibility, and diastolic dysfunction (41–43). These may represent further mechanisms of impaired exercise capacity due to cardiac compression by HH.
Although the present study is the first to evaluate the relationship between cardiac compression and exercise impairment in patients with a large HH, there are several limitations. First, the cohort only included patients with large and symptomatic HHs in whom surgery was already considered appropriate management. Second, we cannot exclude the possibility of a placebo effect of surgery in relation to symptoms and exercise capacity; however, this is unlikely to explain the magnitude of observed improvement in exercise capacity. Studies evaluating intensive exercise training have demonstrated a significantly smaller degree of improved exercise capacity than that seen after HH repair in our patients (44). Third, the absence of an oral load before respiratory investigation may have led to underestimation of the respiratory effect of HH; however, this was consistent for pre-operative and post-operative evaluations. Finally, some of the measures of cardiac compression were subjective; however, there was a good level of interobserver agreement in the evaluation of these parameters, and almost all qualitative evaluations were independently corroborated by quantification.
Patients with large HHs have significant dyspnea and exercise impairment despite normal baseline respiratory function. We show, for the first time, that significant cardiac abnormalities including compression of the left atrium, IPVs, and CS are commonly seen in these patients. The recovery of exercise capacity with HH repair is independently predicted by recovery of the LA diameter, suggesting a significant causal role for cardiac compression in the pathogenesis of HH-associated dyspnea. Assessment of LA compression severity pre-operatively may be a useful noninvasive clinical tool for identifying those patients who will benefit most from HH repair.
The authors thank Prof. Jennifer Peat for her contributions in reviewing the statistical analyses.
For supplemental figures, please see the online version of this article.
Left Atrial Compression and the Mechanism of Exercise Impairment in With Large Hiatal Hernias
All authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- coronary sinus
- computed tomography
- forced expiratory volume in the first second of expiration
- forced vital capacity
- hiatal hernia
- inferior pulmonary vein
- left atrial
- metabolic equivalent
- New York Heart Association
- Received April 13, 2011.
- Revision received June 17, 2011.
- Accepted July 5, 2011.
- American College of Cardiology Foundation
- Zhu J.C.,
- Becerril G.,
- Marasovic K.,
- Ing A.J.,
- Falk G.L.
- Nishimura R.A.,
- Tajik A.J.,
- Schattenberg T.T.,
- Seward J.B.
- Devbhandari M.P.,
- Khan M.A.,
- Hooper T.L.
- Siu C.W.,
- Jim M.H.,
- Ho H.H.,
- et al.
- Chan J.,
- Manning W.J.,
- Appelbaum E.,
- Smith P.,
- Rice K.
- Lam C.S.,
- Borlaug B.A.,
- Kane G.C.,
- et al.
- Anderson B.
- Pauwels R.A.,
- Buist A.S.,
- Calverley P.M.,
- Jenkins C.R.,
- Hurd S.S.
- Pellegrino R.,
- Viegi G.,
- Brusasco V.,
- et al.
- Aduen J.F.,
- Zisman D.A.,
- Mobin S.I.,
- et al.
- Schwartzman D.,
- Lacomis J.,
- Wigginton W.G.
- Lang R.M.,
- Bierig M.,
- Devereux R.B.,
- et al.
- Yock P.G.,
- Popp R.L.
- Miller M.R.,
- Hankinson J.,
- Brusasco V.,
- et al.
- Wanger J.,
- Clausen J.L.,
- Coates A.,
- et al.
- Macintyre N.,
- Crapo R.O.,
- Viegi G.,
- et al.
- Jakeways N.,
- McKeever T.,
- Lewis S.A.,
- Weiss S.T.,
- Britton J.
- Park S.M.,
- Shim C.Y.,
- Choi D.,
- et al.
- Vogel W.M.,
- Apstein C.S.,
- Briggs L.L.,
- Gaasch W.H.,
- Ahn J.
- Kitzman D.W.,
- Brubaker P.H.,
- Morgan T.M.,
- Stewart K.P.,
- Little W.C.