Adult Chest Surgery

Chapter 128. Diaphragm Incisions 

The fan-shaped muscle of the diaphragm arises from the internal circumference of the thorax, with attachments to the sternum, the lower six or seven ribs, and the lumbar vertebral bodies. The muscle fibers also attach posteriorly to the aponeurotic arch of the ligamentum arcuatum externum, which overrides the psoas and quadratus lumborum muscles. Laterally, the fibers of the diaphragm interdigitate with slips from the transversalis muscle of the abdomen to originate from the ribs.1 The right crus is larger and longer than the left and arises from the bodies of the upper three or four lumbar vertebrae. The left crus arises from the upper two lumbar vertebral bodies.

There are three natural openings within the diaphragm (Fig. 128-1). The aortic opening is the most posterior of the three and is formed from fibers comprising the right and left diaphragmatic crura.1 This tunnel is actually behind the diaphragm, not within it, and contains the aorta, azygos vein, and thoracic duct. The esophageal hiatus is slightly more ventral in relation to the aortic hiatus and consists of fibers passing between the aorta and the esophagus toward the right crus, as well as fibers converging on the pericardial tendon. The opening of the inferior vena cava lies within the confluence of the tendons of the right hemithorax and the tendon beneath the pericardium.

Figure 128-1.


Abdominal surface of the diaphragm with three natural openings.


During inspiration, the first rib is elevated and fixed by the scalene muscles of the neck (Fig. 128-2). The external intercostal muscles raise, in turn, each of the lower ribs. Raising these ribs, like a bucket handle that is attached to the sternum and vertebral column, enlarges the thorax and creates the negative pressure that ventilates the lung1 (Fig. 128-3).

Figure 128-2.


Relationship of the diaphragm to rib cage.


Figure 128-3.


"Bucket handle" movement of ribs on inspiration increases the thoracic space and negative thoracic pressure.


The diaphragm is the major muscle of inspiration.2 In the resting state, the central tendon is displaced cephalad into the thorax by the positive intraabdominal pressure. During contraction, the radial muscle fibers pull the tendon down toward the abdominal cavity like a drumhead. This further augments the negative intrathoracic pressure and increases the positive pressure in the abdomen. The diaphragmatic crus contributes to the magnitude of displacement of the central tendon (Fig. 128-4). In fact, if there were only circumferential attachments to the rib cage, the diaphragm would be limited in its ability to displace the lower ribs and enlarge the thoracic cavity. The thicker fascicles of the crus, which lie at a 45- to 90-degree angle to the plane of the fan, pull on the anchoring lumbar spine like a lever and thus fix the central tendon in place. In descending, the fan of the diaphragm displaces the intraabdominal viscera, which do not yield completely because they are bolstered by the anterior abdominal wall. The central tendon becomes a fixed point, from which the radial muscles of the fan contract and thus are able to elevate the lower ribs. Even though the points of muscle attachment to the lumbar spine are more caudal than the attachments to the rib cage, the domed tendon location acts as an attachment point cephalad to the attachment to the rib cage. In fact, contraction of the diaphragm can only raise the lower ribs if the intraabdominal viscera are in situ, and not if the organs have been removed.1 Injuries to the crus have a greater disproportionate effect on the respiratory function of the ipsilateral diaphragm than a similar injury would have to the peripheral muscle.

Figure 128-4.


The crus lies at an angle to the diaphragm muscle.

A forced inspiration will descend the central tendon from one to two rib interspaces. Under normal respiration, each hemidiaphragm provides between 15% and 25% of respiratory muscle function, whereas each side of the combined intercostal muscles provide the remaining percentage.3,4 Under strained respiration, however, the diaphragm can increase its workload to provide up to 80% of the work of breathing.

Pleural and Peritoneal Attachments

The pleurae are tightly adherent to the top surface of the diaphragmatic central tendon and most of the musculature. It is impossible to separate the pleura from the central tendon of each hemidiaphragm. As each pleura curves off the chest wall and folds back on itself on the surface of the diaphragm, there is a circumferential diaphragmatic recess of approximately 1 cm that does not contain pleura5 (Fig. 128-5).

Figure 128-5.


The pleurae do not extend all the way into the diaphragmatic recess.


The peritoneum is less adherent to the undersurface of the diaphragm and can be bluntly mobilized off the diaphragm during extraperitoneal approaches to the abdominal aorta. The plane of dissection lies between the inferior phrenic artery and vein on the muscle side and the peritoneal membrane. The peritoneum separates from the central tendon of the right diaphragm to form the falciform ligament and produces an area directly under the central tendon that does not have peritoneal covering. This is known as the bare area (Fig. 128-6).

Figure 128-6.


The peritoneum does not adhere to the abdominal surface of the right diaphragm where it separates from the central tendon, creating a bare area.

Phrenic Artery and Vein

The superior phrenic arteries are located on the thoracic surface of the diaphragm (Fig. 128-7). These represent small branches from the lower thoracic aorta and traverse the posterior diaphragm over the top portion of each crus close to the mediastinum.1 They terminate in small anastomoses with the musculophrenic and pericardiophrenic arteries, which are both branches from the internal mammary artery. These latter two arteries also supply blood to the phrenic nerve and the pericardial fat pad.6

Figure 128-7.


Thoracic surface of the diaphragm with phrenic artery anatomy.


The inferior phrenic arteries lie on the undersurface of the crus and the dome of the diaphragm (Fig. 128-8). These are small, paired vessels with frequent anatomic variations. They can originate separately from the aorta above the celiac artery. Alternatively, a common trunk arising from either the aorta or the celiac artery gives rise to these two arteries. Occasionally, one vessel will originate from the aorta, whereas the other emerges from one of the renal arteries. Diverging near the crus, the inferior phrenic arteries then course obliquely superior and lateral along the inferior surface of the diaphragm. The left phrenic artery passes posterior to the esophagus and then runs anteriorly along the lateral side of the esophageal hiatus. The right inferior phrenic artery passes behind the inferior vena cava.1

Figure 128-8.


Course of the phrenic artery from the abdominal surface of the diaphragm.

Close to the posterior aspect of the central tendon, both the left and right inferior phrenic arteries divide into a medial and lateral branch. The medial branch extends anteriorly, close to the mediastinum. Branches of this vessel traverse the muscular portion of the diaphragm to anastomose with the musculophrenic and pericardiophrenic arteries. The lateral branch of the inferior phrenic artery courses laterally and forms anastomoses with the lower intercostal arteries. The left inferior phrenic artery provides a minor contribution to the blood supply of the lower esophagus. Both the right and left inferior phrenic arteries have branches to the ipsilateral suprarenal gland. These branches are called the right and left superior suprarenal arteries.1

In general, the venous anatomy in this region parallels that of the arteries. The superior phrenic veins are small and drain anteriorly to the internal mammary vein. The much larger inferior phrenic veins parallel the course of the inferior phrenic arteries. The right vein empties directly into the inferior vena cava. The left vein usually has two branches, one of which drains into the left renal or suprarenal vein and the other of which passes anterior to the esophageal hiatus and empties into the inferior vena cava.1

Diaphragmatic Innervation

The phrenic nerve originates from the C3, C4, and C5 nerve roots and then enters the chest anterior to the subclavian artery. On the left-hand side, the nerve lies medial to the internal mammary (thoracic) artery 64% of the time, and on the right-hand side, it lies medial to the internal mammary artery only 46% of the time7 (Fig. 128-9). Thus the left nerve is more prone to injury during mobilization of the left internal mammary artery through a median sternotomy incision.

Figure 128-9.


Juxtaposition of the phrenic nerve and internal thoracic (mammary) artery on right and left sides creates the potential for phrenic nerve injury, leading to diaphragmatic elevation.


In utero, the phrenic nerve elongates as the septum transversum migrates caudally. There is, however, an outer rim of diaphragmatic muscle that originates from migrating mesenchymal cells of the nearby body wall innervated by spinal nerves from thoracic levels T7-T12. Additionally, there is a contribution from mesenchyme associated with the foregut at levels L1-L3 that coalesces to form the right and left crura.3

Despite the contributions from the thoracic and spinal nerve roots, the majority of the diaphragm is innervated by the phrenic nerve. While the origin of the phrenic nerve and its proximal course through the mediastinum are well known, the distal extent of the nerve as it branches into the diaphragm proper is less well described. In 1956, Merendino and colleagues published the most descriptive and often-cited reference regarding this intradiaphragmatic portion of the phrenic nerve.8 Their anatomic findings and frequently adapted schematized drawings are based on electrical stimulation studies and gross dissection in dogs as well as intraoperative dissection of approximately 40 human diaphragms.

The phrenic nerve usually divides at the level of the diaphragm or just above it. The right phrenic nerve enters the diaphragm just lateral to the inferior vena cava within the central tendon (Fig. 128-10). The left phrenic nerve enters lateral to the left border of the heart just anterior to the central tendon within the muscle itself. The intradiaphragmatic course of the phrenic nerve can be predicted, even when not directly seen, by knowing the distribution of the four main motor divisions. The phrenic nerve first splits into an anterior and posterior trunk. The anterior trunk subsequently divides into a sternal and anterolateral branch near the anteromedial border of the central tendon. The posterior trunk likewise divides into a crural and posterolateral branch along the posteromedial border of the central tendon. The sternal and crural branches are short and continue to run in an anteromedial and posteromedial direction, respectively. The anterolateral and posterolateral branches are much longer and run close to the muscular fiber insertions into the central tendon. These two branches innervate the majority of the diaphragm. Their anatomic relation to one another is often described as a pair of handcuffs or manacles. Often these branches are within the muscle layers and are not readily visible.

Figure 128-10.


The phrenic nerves pierce the diaphragmatic surface and course through musculature invisible to the naked eye.


Diaphragmatic incisions can be divided into three groups: circumferential, central tendon, and radial.


Circumferential incisions in the periphery result in little loss of function. These circumferential incisions, however, must lie at least 5 cm lateral to the edge of the central tendon to avoid the posterolateral and anterolateral branches of the phrenic nerve. These incisions can be difficult to correctly realign after a long operation. Placing surgical clips on each side of the muscular incision can greatly facilitate the correct spatial orientation on closing (Fig. 128-11).

Figure 128-11.


Circumferential incision.

Central Tendon

Incisions in the central tendon, as far centrally as within 2 cm of the entrance of the phrenic nerve, do not interrupt any major branch of the nerve itself. This type of incision can provide excellent visualization of the abdomen from the thorax, and vice versa. These incisions are easy to open and close (Fig. 128-12).

Figure 128-12.


Central tendon incision.


A transverse radial incision made from the midaxillary line centrally is relatively safe because it courses between the distal aspects of the anterolateral and posterolateral branches of the phrenic nerve (i.e., through the opening of the "handcuffs") (Fig. 128-13). Radial incisions from the costal margin extending all the way to the esophageal hiatus, however, may result in segmental diaphragmatic paralysis if the incision transects the crural or posterolateral branches of the phrenic nerve.

Figure 128-13.


Radial incision.


The two most common causes of diaphragmatic elevation are congenital eventration of the diaphragm and phrenic nerve palsy9,10 (see Chap. 129). Phrenic nerve paralysis can occur as a consequence of a viral palsy, iatrogenic injury during thoracic surgery, or fracture of an anterior first rib or clavicle.

Plication for Phrenic Nerve Palsy

The easiest technique of plication is to place imbricating stitches within the central tendon of the diaphragm.11,12 These stitches can be readily placed both thoracoscopically and through a minithoracotomy.13 If the sutures extend far enough from the edge of the diaphragmatic tendon, they can produce substantial caudal displacement of the tendon toward the abdominal cavity, permitting expansion of the ipsilateral lower lobe as well as balancing the mediastinum14 (Fig. 128-14). The drawback of this technique, however, is that while the majority of the pleats will fold the noncompliant central tendon, the compliant muscular remnant can be expected to stretch with time. It has been our experience that the central tendon pleating technique is associated with reelevation of the diaphragm to the level of the hilum within several years. Long-term follow-up of this technique has included reports of recurrent diaphragmatic elevation requiring additional intervention in as many as 19% of treated patients.15 Furthermore, the phrenic vessels and branches of the nerve travel near the insertion of the muscle into the edge of the tendon and cannot be visualized from the thoracic surface of the diaphragm. Yet, to displace the central tendon adequately, these stitches need to extend into the muscle area, which places the branches of the nerve at risk of injury.

Figure 128-14.


Central tendon plication technique.


The top (thoracic) view of the diaphragm seen in Fig. 128-10 illustrates the radial spokes of the muscle fibers from their origin along the costal margin toward the central tendon. The phrenic nerve can be seen along the mediastinal pleura, but then it pierces the diaphragmatic muscle close to the inferior vena cava on the right and the tip of the acute angle of the heart on the left. The phrenic vessels and phrenic nerve cannot be visualized from the thoracic surface of the diaphragm beyond these areas.

Dr. David State described a subcostal radial plication technique for congenital eventration of the diaphragm in 1949.16 The original description of this technique included a generous incision across the right upper quadrant of the abdomen and placement of radial sutures along the muscular portion of the diaphragm to pull it toward the lateral chest wall. A transthoracic radial plication also has been described.14,17

Dr. John Foker at the University of Minnesota has used a transthoracic radial plication technique since 1976 to treat 35 children with elevation of the diaphragm.18 The repairs were performed with interrupted horizontal mattress pledgeted sutures imbricating the muscular portion of the diaphragm in a radial manner toward the chest wall via a posterolateral thoracotomy (Fig. 128-15). The plication sutures extend in an unbroken band from the xiphoid area to the vertebral body. No sutures are placed along the mediastinal pleura. The goal is to produce a taut diaphragm that appears as a straight, angled line from mediastinum to chest wall on an anteroposterior roentgenogram of the chest. I believe that this produces a plication that best mimics the contraction of the fan-shaped muscle while minimizing injury to the branches of the nerve or vessels.

Figure 128-15.


Radial plication technique.

In this series, 31 of the 36 operations (86%) led to extubation within 3 days, even though 15 patients had been ventilator-dependent before plication.18 There were no deaths within 30 days and no morbidity directly attributed to the plication. Only one patient (3%) suffered a recurrence requiring repeat plication. Twenty-six of these patients survived long term (median 12 years at time of analysis), and eighteen of these patients were reevaluated with diaphragmatic ultrasound in 1996. Some degree of function had returned to 14 (78%) of the diaphragms.

I have extended this technique to a thoracoscopic approach in adults with elevated hemidiaphragms with some success. Currently, I use a three-port technique with an anterior and posterior port at the sixth and eighth intercostal spaces, respectively. The third port is subcostal and is used to pass an O-ring clamp through the abdominal cavity to grasp the undersurface of the central tendon of the diaphragm. This permits vigorous caudal displacement of the muscle to visualize the muscular imbrications for plication. The posterior thoracic port then is used to plicate the anterior and lateral borders of the muscle, while the anterior thoracic port is used to plicate the lateral and posterior borders.


Partial diaphragmatic resections are necessary to remove tumors that have invaded a portion of the diaphragm. The redundancy of the muscle frequently permits primary repair for small to modest resections. Larger defects are easily repaired with the use of a mesh or impermeable graft sutured to the remnant of muscle. I use mesh for patients with lung tissue remaining in the ipsilateral hemithorax and impermeable grafts for patients who have had a pneumonectomy to prevent fluid shifts between the thorax and abdomen.

Complete diaphragmatic resection may be required for large tumors invading the diaphragm, such as lung cancer of the lower lobe or sarcomas of the chest. I have gained extensive experience with complete diaphragmatic resection in the course of developing the extrapleural pneumonectomy for mesothelioma.5 Full details of patch replacement of the entire hemidiaphragm appear in Chapter 103.


The diaphragm serves several functions as a result of its unique anatomic location. It is the major muscle of respiration by creating the intrathoracic vacuum and moving the lower ribs. It is a barrier between the positive-pressure abdomen and the negative-pressure thorax.

Surgical approaches to the diaphragm depend on an understanding of the anatomy and physiology of the muscle, nerve, and blood vessels. A number of incisions are possible once the location of the nerve and vessel branches have been learned. These structures frequently lie within the muscle itself and are not seen on the surface of the structure. Therefore, the concept of a nerve "manacle" around the junction of the central tendon to the muscle is a very helpful visual mnemonic.

Diaphragmatic plication is helpful in acquired injuries to the phrenic nerve. I prefer a radial plication technique for phrenic nerve palsy that mimics the radial contraction of the fan-shaped muscle to that of a central tendon pleat.

Although the diaphragm acts as a boundary between the thorax and abdomen, it should not serve as a boundary between the realm of thoracic surgeons and gastrointestinal surgeons. The diaphragm makes a good fence, but neighbors on both sides of that fence should know its anatomy, physiology, and surgical principles of resection and repair.


A 60-year-old man with baseline chronic obstructive pulmonary disease presented with a viral left phrenic nerve palsy. He had developed a viral pleurisy of the left chest about 1 month previously, with sudden progression of his breathing dysfunction. On presentation, he was dyspneic on modest exertion. He also was suffering from severe postprandial bloating of the left upper quadrant and could not belch. Upright chest x-ray showed elevation of the left diaphragm to the level of the left hilum without mediastinal adenopathy. An ultrasound evaluation of the diaphragm revealed paradoxical motion on the left consistent with phrenic nerve palsy. Further radiographic workup demonstrated no evidence of malignancy. His pulmonary function tests revealed a forced expiratory volume in 1 second of only 32% of predicted.

Based on his symptoms and depressed pulmonary function, the decision was made to perform a left-sided video-assisted diaphragmatic plication using a radial technique. Postoperative chest x-ray showed a flat and taut diaphragm displaced lower than normal in the thorax. Although his symptoms of gas bloat and dyspnea were markedly improved, neither had completely returned to normal within the first year. Repeat pulmonary function tests showed modest improvement in the forced expiratory volume in 1 second to 41% of predicted.

His shortness of breath was improved within the first week and then continued to improve gradually over the next 6 months. His dyspnea remained fairly stable for over an additional year and then improved dramatically over a period of about 2 months. His chest x-ray also changed, with the diaphragm assuming the normal contour and position. Ultrasound evaluation of the diaphragm showed resolution of the previously noted paradoxical motion. The phrenic nerve had recovered.


Those who do a lot of esophageal or mesothelioma surgery are well versed in the diaphragmatic anatomy and incisions. I tend to divide the diaphragm using a stapler which avoids the use of stay or marking sutures that can get in your way. In addition, the staple line can serve as a buttress as well. In terms of diaphragmatic plications, I find that the radial technique tends to result in lower gains of thoracic volume than the other techniques as a consequence of the more difficult visibility and exposure.



1. Gray H: Anatomy, Descriptive and Surgical. Revised American, from the 15th English ed, T. Pick and R. Howden (eds). 1977, New York: Bounty Books. pp. 1257. 

2. Epstein SK: An overview of respiratory muscle function. Clin Chest Med. 15:619–39, 1994.[PubMed: 7867279]

3. Wade OL: Movements of the thoracic cage and diaphragm in respiration. J Physiol 124:193–212, 1954.[PubMed: 13175123]

4. Bergofsky EH: Relative Contributions of the Rib Cage and the Diaphragm to Ventilation in Man. J Appl Physiol 19:698–706, 1964.[PubMed: 14195582]

5. Sugarbaker DJ, Mentzer SJ, Strauss G: Extrapleural pneumonectomy in the treatment of malignant pleural mesothelioma. Ann Thorac Surg 54:941–6, 1992.[PubMed: 1417290]

6. Anderson TM, Miller JI , Jr.: Surgical technique and application of pericardial fat pad and pericardiophrenic grafts. Ann Thorac Surg 59:1590–1, 1995.[PubMed: 7771856]

7. Owens WA, Gladstone DJ, Heylings DJ: Surgical anatomy of the phrenic nerve and internal mammary artery. Ann Thorac Surg 58:843–4, 1994.[PubMed: 7944713]

8. Merendino KA, Johnson RJ, Skinner HH, Maguire RX: The intradiaphragmatic distribution of the phrenic nerve with particular reference to the placement of diaphragmatic incisions and controlled segmental paralysis.Surgery 39:189–98, 1956.[PubMed: 13298965]

9. Beck WC, Motsay DS: Eventration of the diaphragm. AMA Arch Surg 65:557–63, 1952.[PubMed: 12984955]

10. Marcos JJ, Grover FL, Trinkle JK: Paralyzed diaphragm–effect of plication on respiratory mechanics. J Surg Res 16:523–6, 1974.[PubMed: 4831729]

11. Garbaccio C, Gyepes MT, Fonkalsrud EW: Malfunction of the intact diaphragm in infants and children. Arch Surg 105:57–61, 1972.[PubMed: 5031599]

12. Schwartz MZ, Filler RM: Plication of the diaphragm for symptomatic phrenic nerve paralysis. J Pediatr Surg 13:259–63, 1978.[PubMed: 671190]

13. Mouroux J, Padovani B, Poirier NC, Benchimol D, Bourgeon A, Deslauriers J, Richelme H: Technique for the repair of diaphragmatic eventration. Ann Thorac Surg 62:905–7, 1996.[PubMed: 8784039]

14. Shoemaker R, Palmer G, Brown JW, King H: Aggressive treatment of acquired phrenic nerve paralysis in infants and small children. Ann Thorac Surg 32:250–9, 1981.[PubMed: 7283517]

15. Smith CD, Sade RM, Crawford FA, Othersen HB: Diaphragmatic paralysis and eventration in infants. J Thorac Cardiovasc Surg 91:490–7, 1986.[PubMed: 3959566]

16. State D: The surgical correction of congenital eventration of the diaphragm in infancy. Surgery 25:461–8, 1949.[PubMed: 18113335]

17. Sethi G, Reed WA: Diaphragmatic malfunction in neonates and infants. Diagnosis and treatment. J Thorac Cardiovasc Surg 62:138–43, 1971.[PubMed: 5091677]

18. Jaklitsch M: Twenty-year experience with peripheral radial plication of the diaphragm. 33rd Annual Meeting of the Society of Thoracic Surgeons San Diego, CA:1997.

If you find an error or have any questions, please email us at Thank you!