Secondary Logo

Journal Logo

Regional Anesthesia: Research Report

Ultrasound-Guided Thoracic Paravertebral Blockade

A Retrospective Study of the Incidence of Complications

Pace, Meredith M. MD; Sharma, Balram MD; Anderson-Dam, John MD; Fleischmann, Katherine MD; Warren, Lisa MD; Stefanovich, Peter MD

Author Information
doi: 10.1213/ANE.0000000000001117
  • Free

In 2013, the American Cancer Society estimated the diagnosis of 232,340 new cases of invasive breast cancer and 64,640 additional cases of breast carcinoma in situ in the United States.1 With many of these patients presenting for mastectomy, consideration must be given to anesthetic techniques that include thoracic paravertebral block (TPVB). TPVB produces an ipsilateral somatic and sympathetic nerve blockade that spreads from the site of injection across several contiguous dermatomes.2 This block can effectively control acute postoperative pain, which is a risk factor for the development of chronic pain after breast cancer surgery.3,4

Despite numerous studies showing the clinical efficacy of TPVB, serious complications, such as pneumothorax, have been associated with this type of regional anesthetic. Therefore, widespread use may still be limited by safety concerns. So far, however, complications have been reported only for blocks placed without the guidance of ultrasound.5–9 Multiple authors have suggested that an ultrasound-guided approach may reduce risk.9–11 We hypothesize that the incidence of complications from the single-injection, transverse, in-plane ultrasound-guided approach for TPVB is low in a large cohort of patients undergoing mastectomy with reconstruction for breast cancer treatment or prophylaxis.


An IRB approval was obtained for this study. Patients who underwent TPVB before unilateral or bilateral mastectomy with immediate tissue expander or implant breast reconstruction at Massachusetts General Hospital between January 1, 2010, and December 3, 2013, were identified from a prospectively maintained regional anesthesia database. Patients having sentinel lymph node biopsy or axillary lymph node dissection, in addition to mastectomy with breast reconstruction, were included in the study. Patient age, sex, body mass index, type of surgery, laterality of block performed, local anesthetic type, concentration, and volume used were entered into the database at the time of block placement. Any procedural difficulties encountered, to include suspected accidental pleural puncture, were documented. Symptoms of local anesthetic toxicity exhibited by the patient were recorded, as well as any subsequent intervention. Hemodynamic instability after block placement was documented, as well as a description of interventions. Because most TPVBs were performed in a short time before transport to the operating room, a detailed evaluation of sensory block distribution was not performed and recorded for every patient. On the first postoperative day, a member of the regional anesthesia team reviewed each patient’s intraoperative anesthesia, postanesthesia care, and inpatient records. The records were specifically screened for any evaluation and treatment of pneumothorax, as well as documentation of lower extremity or contralateral upper extremity sensory or motor deficit suggestive of epidural or spinal spread of local anesthetic. Each patient was also interviewed and examined by a member of the regional anesthesia team postoperatively to assess block duration and overall satisfaction with the regional anesthetic. Study staff reviewed medical records from any patient for whom a complication related to the regional anesthetic was recorded in the database. For these cases, nursing records were reviewed carefully for the changes in vital signs relative to the block procedure. Any interventions that took place, as well as the overall clinical outcome of the patient, were noted.

All patients had a unilateral or bilateral TPVB placed preoperatively for the purpose of providing postoperative analgesia. A dedicated block service resident or fellow, along with an attending regional anesthesiologist, were involved in the placement of all blocks. Procedures were performed with patients in the prone position after IV access was secured. Standard monitoring was used in all cases, including noninvasive arterial blood pressure, pulse oximetry, and 5-lead continuous electrocardiogram. Supplemental oxygen was provided through nasal cannula. Patients were sedated with 0 to 4 mg midazolam and 0 to 200 μg fentanyl, delivered IV, as needed. Skin of the upper back was prepared with 2% chlorhexidine. A 5- to 15-MHz linear array ultrasound transducer probe (General Electric, Logiq e, Wauwatosa, WI) placed in a sterile sleeve was used to visualize the paravertebral space at the T3 or T4 level. A skin wheal was created using a 25-gauge needle to inject 1 to 2 mL 1% lidocaine.

Transverse, In-Plane, Ultrasound-Guided Block Technique

An appropriate thoracic spinous process was located by positioning the probe in the transverse plane. The probe was then moved laterally to locate the transverse process. The probe was manipulated slightly caudad or cephalad to locate the intercostal space and avoid acoustic shadowing from neighboring ribs. The transverse process was visualized medially with the pleura dipping under the inferolateral aspect (Fig. 1). The internal intercostal membrane, which is contiguous with the superior costotransverse ligament, was generally seen as a thin, radiopaque line extending from the transverse process, creating a wedge-shaped pocket, which represents the thoracic paravertebral space. A 22-gauge, facet-tip needle (SonoPlex, Pajunk Medical Systems, L.P., Norcross, GA) was advanced, in plane, from the lateral aspect of the ultrasound probe (Fig. 2). When the needle pierced the internal intercostal membrane, and after aspiration demonstrated the absence of air or blood, 15 to 20 mL 0.5% bupivacaine with 1:400,000 epinephrine was deposited in 5-mL increments. Twenty milliliters of local anesthetic was injected for unilateral blocks and a total of 30 mL (15 mL on each side) for bilateral procedures. Depression of the pleura was considered the end point of the procedure, and this was clearly visualized in every case (Fig. 3).

Figure 1
Figure 1:
Ultrasound image of the thoracic paravertebral space at the level of T3 to T4. IICM = internal intercostal membrane; PL = pleura; PVS = apex of the thoracic paravertebral space; TP = transverse process.
Figure 2
Figure 2:
Ultrasound image of in-plane needle advancement into the thoracic paravertebral space. The needle tip can be visualized on the dorsal side of the internal intercostal membrane. N = block needle; PL = pleura; TP = transverse process.
Figure 3
Figure 3:
Ultrasound image of local anesthetic injection into the thoracic paravertebral space. Depression of the medial aspect of the pleura can be observed just beneath the needle. LA = local anesthetic; N = block needle; PL = pleura; TP = transverse process.

Patients’ vital signs were monitored continuously during block placement and for a minimum of 30 minutes afterward. All patients underwent general anesthesia for their surgical procedures and were admitted to the hospital for at least 1 postoperative night.

Patient demographics and complication rates among the unilateral and bilateral TPVB patients were compared. For continuous variables, the 2-sample t test was used. For categorical variables (American Society of Anesthesiologists physical status), the Mann-Whitney U test was used. Statistical significance was declared at α = 0.05.

Confidence intervals (CIs) for proportions were calculated using the Clopper-Pearson method for the binomial distribution.12 For outcomes with an incidence of 0, the rule of 3 calculation was used to estimate the CI. The exact test was used to compare the incidence of complications between patients who received unilateral and bilateral TPVB. Statistical significance of the relative risk was assessed using the χ2 test.


During the time period observed, 856 patients underwent TPVB before undergoing mastectomy with immediate breast reconstruction. Two hundred eighty-five unilateral and 571 bilateral blocks were performed, for a total of 1427 paravertebral injections. All patients were female and received 0.5% bupivacaine with 1:400,000 epinephrine as the local anesthetic for TPVB. Patient demographic characteristics were similar, with no statistically significant differences observed between those patients who received unilateral versus bilateral blocks (Table 1).

Table 1
Table 1:
Patient Characteristics

Six major complications related to TPVB placement among 856 patients were recorded in the database (0.70%; 99.2% CI, 0.17%–1.86%; Fig. 4). The incidence of both suspected accidental pleural puncture and symptomatic pneumothorax was 0 in 856 patients (99.2% CI, 0.00%–0.35%). The incidence of symptomatic hypotension and bradycardia was 4 in 856 patients (0.47%; 99.2% CI, 0.07%–1.50%). The incidence of suspected toxicity from local anesthetic was 2 in 856 patients (0.23%; 99.2% CI, 0.01%–1.11%). The incidence of symptoms suggestive of local anesthetic solution spread outside of the paravertebral space was 0 in 856 patients (99.2% CI, 0.00%–0.35%). The incidence of complications for patients who received unilateral TPVB was 0.35% (99.2% CI, 0.00%–2.66%). The incidence of complications for patients who received bilateral TPVB was 0.88% (99.2% CI, 0.18%–2.51%; Table 2).

Table 2
Table 2:
Complications from TPVB
Figure 4
Figure 4:
Major complications from thoracic paravertebral block placement.

Three of the 4 patients whose cases were complicated by symptomatic hemodynamic changes (cases 1, 4, and 5) developed a decrease in heart rate and arterial blood pressure within 15 minutes of block procedure completion. All 3 of these cases involved patients who had received bilateral TPVB. In another case (case 3), the patient complained of feeling unwell 15 minutes after the completion of a unilateral TPVB. She developed progressively worsening bradycardia that was rapidly followed by loss of consciousness and a brief period of asystole. The advanced cardiac life support protocol was initiated, and the patient regained cardiovascular stability after 12 seconds of chest compressions and 1 mg IV atropine. On consultation with cardiology, a vasovagal episode precipitated by hemodynamic changes that occurred secondary to the TPVB was felt to be the most likely etiology of the event.

There were 2 cases of possible local anesthetic toxicity (cases 2 and 6); 1 patient was given IV 20% lipid emulsion and midazolam prophylactically for symptoms suggestive of central nervous system toxicity. There was no evidence of cardiovascular toxicity in either case. In all 6 cases, surgery was completed as planned on the day of block placement.


By using ultrasound technology, the anatomy of the paravertebral space may be clearly defined in each patient. Needle advancement can be guided by real-time imaging, thereby potentially improving safety for this particular block technique. In 2009, Shibata and Nishiwaki13 first described an ultrasound-guided TPVB using a transverse probe position and an in-plane view of the needle. We report the incidence of complications from TPVB using this technique in a large cohort of patients undergoing mastectomy with reconstruction for breast cancer treatment or prophylaxis.

In 2001, Naja and Lönnqvist7 reported the incidence of complications from thoracic and lumbar paravertebral blocks placed using a landmark technique facilitated by nerve stimulation. By using this block technique, the overall incidence of pleural puncture was 0.8%, and the subsequent development of pneumothorax was 0.5%. Pleural puncture was identified by a typical “pop” felt by the proceduralist, followed by an unusually easy injection of local anesthetic. Pneumothorax was diagnosed by chest radiograph in those patients who had experienced pleural puncture. In a smaller study by Terheggen et al.,14 the loss-of-resistance technique was used for TPVB in 15 patients, 1 of whom experienced an accidental pleural puncture that was suspected after a greater than normal loss of resistance to local anesthetic injection. Incidentally, radiographic follow-up in this case did not identify a pneumothorax. Our group did not identify any suspected accidental pleural puncture or encounter any patient who developed clinical symptoms of a pneumothorax after TPVB placement. Avoidance of these complications may be a consequence of improved safety using real-time needle visualization and imaging of the pleura with ultrasound. Indeed, there is the possibility that there may have been patients in our cohort in whom intrapleural injection or pneumothorax did occur, but this was not identified because of lack of clinical suspicion in conjunction with the absence of patient symptoms.

Four cases of symptomatic hemodynamic changes were identified in this study, 3 of which involved a decrease in arterial blood pressure and heart rate resulting in symptoms of lightheadedness or nausea after block placement. In a fourth case, the patient experienced what was most likely a vasovagal episode 15 minutes after the completion of a unilateral TPVB. Hemodynamic changes that occurred as a consequence of the TPVB were felt to be the most likely etiology of the event. The time from block completion to the onset of signs and symptoms was between 10 and 15 minutes in all 4 cases. Sympathetic blockade that can be produced by TPVB2,15 may unmask hypovolemia and explain cases of symptomatic hemodynamic instability.

We also identified 2 cases of suspected local anesthetic toxicity in our study cohort. Incremental aspiration is routinely performed during local anesthetic injection, so the likelihood of a direct intravascular injection in these 2 cases is low. Therefore, the signs and symptoms seen in these patients are more likely explained by the absorption of bupivacaine from the paravertebral or communicating epidural and intercostal spaces.16 More rapid absorption of local anesthetic, as from highly vascularized anatomic locations such as those involved with the paravertebral block, is associated with a higher incidence of toxicity.17 Berrisford et al.18 examined the time to peak plasma concentration of bupivacaine in adult patients who had received a 20-mL bolus dose of 0.5% bupivacaine through an intercostal catheter placed under direct visualization during thoracotomy. The average time to peak plasma concentration in this investigation was 25 minutes. On the basis of these findings, patients who received TPVB at our institution were routinely monitored for a minimum of 30 minutes after procedure completion, regardless of the total time needed to place the block. Patients who needed to be transported to the operating room before the end of this designated time period continued to be monitored en route with portable equipment. Our choice of local anesthetic and volume is based on the previous studies where administration of 50 to 150 mg of 0.5% bupivacaine with epinephrine is described.13,19,20 However, there are no published data regarding optimal dose and concentration of local anesthetic for single-shot TPVB. Given that the paravertebral and communicating spaces are highly vascularized, it seems reasonable to consider the use of a more dilute local anesthetic to decrease the risk of toxicity from systemic absorption. Moreover, larger volumes of injectate may not necessarily increase the anesthetic effect of TPVB, because the distribution of local anesthetic has been shown to be highly variable.16

Symmetrical contralateral anesthesia has been described as a consequence of accidental epidural,21 intrathecal,22,23 and subdural24 spread of local anesthetic during attempted paravertebral space cannulation. Moreover, it has been suggested that ultrasound-guided approaches to the paravertebral space with a lateral-to-medial needle direction are associated with more frequent epidural and intrathecal spread.25 In our study cohort, there was no reported incidence of contralateral anesthesia after unilateral TPVB. However, general anesthesia was induced shortly after placement of the block in many cases, not allowing time for a comprehensive evaluation of sensory blockade. In addition, the average time under anesthesia was 3.9 hours for unilateral surgical procedures and 5.4 hours for bilateral cases. If a complication related to local anesthetic spread outside of the thoracic paravertebral space occurred, transient effects might have been masked.

In conclusion, our data suggest that a transverse, in-plane ultrasound-guided TPVB has a low risk of complications. The lack of suspected accidental pleural puncture and symptomatic pneumothorax in the current cohort of patients might be attributable to the use of ultrasound guidance.


Name: Meredith M. Pace, MD.

Contribution: This author helped design the study, collect the data, analyze the data, and prepare the manuscript.

Attestation: Meredith M. Pace approved the final manuscript and attests to the integrity of the original data and the analysis reported in this manuscript.

Name: Balram Sharma, MD.

Contribution: This author helped design the study, collect the data, analyze the data, and prepare the manuscript.

Attestation: Balram Sharma approved the final manuscript and attests to the integrity of the original data and the analysis reported in this manuscript.

Name: John Anderson-Dam, MD.

Contribution: This author helped design the study, collect the data, analyze the data, and prepare the manuscript.

Attestation: John Anderson-Dam approved the final manuscript and attests to the integrity of the original data and the analysis reported in this manuscript.

Name: Katherine Fleischmann, MD.

Contribution: This author helped collect the data and prepare the manuscript.

Attestation: Katherine Fleischmann approved the final manuscript and attests to the integrity of the original data and the analysis reported in this manuscript.

Name: Lisa Warren, MD.

Contribution: This author helped collect the data and prepare the manuscript.

Attestation: Lisa Warren approved the final manuscript and attests to the integrity of the original data and the analysis reported in this manuscript.

Name: Peter Stefanovich, MD.

Contribution: This author helped design the study, collect the data, analyze the data, and prepare the manuscript.

Attestation: Peter Stefanovich approved the final manuscript and attests to the integrity of the original data and the analysis reported in this manuscript and is the archival author.

This manuscript was handled by: Terese T. Horlocker, MD.


The authors thank the following individuals for their contributions: Kwun Yee Trudy Poon, Biostatistician and Epidemiologist, and Katherine Koury, Research Coordinator (Department of Anesthesia, Critical Care and Pain Medicine, Massachusetts General Hospital, Boston, Massachusetts).


1. American Cancer Society. Breast Cancer Facts & Figures 2013–2014. 2013 Atlanta, GA American Cancer Society, Inc
2. Andreae MH, Andreae DA. Regional anaesthesia to prevent chronic pain after surgery: a Cochrane systematic review and meta-analysis. Br J Anaesth. 2013;111:711–20
3. Cheema SP, Ilsley D, Richardson J, Sabanathan S. A thermographic study of paravertebral analgesia. Anaesthesia. 1995;50:118–21
4. Gärtner R, Jensen MB, Nielsen J, Ewertz M, Kroman N, Kehlet H. Prevalence of and factors associated with persistent pain following breast cancer surgery. JAMA. 2009;302:1985–92
5. Schnabel A, Reichl SU, Kranke P, Pogatzki-Zahn EM, Zahn PK. Efficacy and safety of paravertebral blocks in breast surgery: a meta-analysis of randomized controlled trials. Br J Anaesth. 2010;105:842–52
6. Purcell-Jones G, Pither CE, Justins DM. Paravertebral somatic nerve block: a clinical, radiographic, and computed tomographic study in chronic pain patients. Anesth Analg. 1989;68:32–9
7. Naja Z, Lönnqvist PA. Somatic paravertebral nerve blockade. Incidence of failed block and complications. Anaesthesia. 2001;56:1184–8
8. Richardson J, Lönnqvist PA, Naja Z. Bilateral thoracic paravertebral block: potential and practice. Br J Anaesth. 2011;106:164–71
9. Cowie B, McGlade D, Ivanusic J, Barrington MJ. Ultrasound-guided thoracic paravertebral blockade: a cadaveric study. Anesth Analg. 2010;110:1735–9
10. Ben-Ari A, Moreno M, Chelly JE, Bigeleisen PE. Ultrasound-guided paravertebral block using an intercostal approach. Anesth Analg. 2009;109:1691–4
11. O Riain SC, Donnell BO, Cuffe T, Harmon DC, Fraher JP, Shorten G. Thoracic paravertebral block using real-time ultrasound guidance. Anesth Analg. 2010;110:248–51
12. Clopper CJ, Pearson ES. The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika. 1934;26:404–13
13. Shibata Y, Nishiwaki K. Ultrasound-guided intercostal approach to thoracic paravertebral block. Anesth Analg. 2009;109:996–7
14. Terheggen MA, Frank W, Borel Rinks IH, Ionescu TI, Knape JT. Paravertebral blockade for minor breast surgery. Anesth Analg. 2002;94:355–9
15. Richardson J, Jones J, Atkinson R. The effect of thoracic paravertebral blockade on intercostal somatosensory evoked potentials. Anesth Analg. 1998;87:373–6
16. Marhofer D, Marhofer P, Kettner SC, Fleischmann E, Prayer D, Schernthaner M, Lackner E, Willschke H, Schwetz P, Zeitlinger M. Magnetic resonance imaging analysis of the spread of local anesthetic solution after ultrasound-guided lateral thoracic paravertebral blockade: a volunteer study. Anesthesiology. 2013;118:1106–12
17. Rosenberg PH, Veering BT, Urmey WF. Maximum recommended doses of local anesthetics: a multifactorial concept. Reg Anesth Pain Med. 2004;29:564–75
18. Berrisford RG, Sabanathan S, Mearns AJ, Clarke BJ, Hamdi A. Plasma concentrations of bupivacaine and its enantiomers during continuous extrapleural intercostal nerve block. Br J Anaesth. 1993;70:201–4
19. Pusch F, Freitag H, Weinstabl C, Obwegeser R, Huber E, Wildling E. Single-injection paravertebral block compared to general anaesthesia in breast surgery. Acta Anaesthesiol Scand. 1999;43:770–4
20. Cheema S, Richardson J, McGurgan P. Factors affecting the spread of bupivacaine in the adult thoracic paravertebral space. Anaesthesia. 2003;58:684–7
21. Frohm RM, Raw RM, Haider N, Boezaart AP. Epidural spread after continuous cervical paravertebral block: a case report. Reg Anesth Pain Med. 2006;31:279–81
22. Sharrock NE. Postural headache following thoracic somatic paravertebral nerve block. Anesthesiology. 1980;52:360–2
23. Evans PJ, Lloyd JW, Wood GJ. Accidental intrathecal injection of bupivacaine and dextran. Anaesthesia. 1981;36:685–7
24. Garutti I, Hervias M, Barrio JM, Fortea F, De La Torre J. Subdural spread of local anesthetic agent following thoracic paravertebral block and cannulation. Anesthesiology. 2003;98:1005–7
25. Chelly JE. Paravertebral blocks. Anesthesiol Clin. 2012;30:75–90
© 2016 International Anesthesia Research Society