Secondary Logo

Journal Logo

Pupillary Reflex for Evaluation of Thoracic Paravertebral Block: A Prospective Observational Feasibility Study

Duceau, Baptiste MD*; Baubillier, Mélanie MD*; Bouroche, Gaëlle MD; Albi-Feldzer, Aline MD*; Jayr, Christian MD, PhD*

doi: 10.1213/ANE.0000000000002003
Regional Anesthesia and Acute Pain Medicine: Original Clinical Research Report

BACKGROUND: Although thoracic paravertebral block (TPVB) is recommended in major breast surgery, there is no gold standard to assess the success of TPVB. Pupillary dilation reflex (PDR) is the variation of the pupillary diameter after a noxious stimulus. The objective was to evaluate the feasibility of recording the PDR to assess analgesia in an anesthetized thoracic dermatome after TPVB.

METHODS: This prospective, observational, single-center study included 32 patients requiring breast surgery under general anesthesia and TPVB. TPVB was performed before surgery under ultrasound guidance with 20 mL of 0.75% ropivacaine. At the end of the surgery, remifentanil was stopped and the PDR was recorded after a 5-second tetanic stimulation (60 mA, 100 Hz) applied to the anterior chest wall. The PDR was defined as the maximal increase in pupil diameter after a standardized noxious stimulus, expressed as a percentage of the initial pupil diameter. The PDR was recorded twice in the same eye for each patient after a stimulus on both the TPVB and the control sides. Postoperative pain scores were recorded in a postanesthesia care unit. The primary outcome was the difference between the PDR on the TPVB and the control sides.

RESULTS: The median (interquartile range) PDR was 9% (4%–13%) on the TPVB side and 41% (27%–66%) on the control side. There was a significant difference in the PDR between the TPVB and the control sides with a Hodges-Lehmann estimate of absolute difference of 37% points (95% confidence interval, 25–52, P < .001). Median postoperative pain scores (interquartile range) in the postanesthesia care unit were 1 (0–3) at rest and 1 (0–3) during mobilization, respectively. There was a linear correlation between maximal postoperative pain scores and the PDR on the TPVB side with a Pearson’s correlation coefficient r = 0.40 (95% confidence interval, 0.06–0.66, P = .02). No correlation was found between the number of blocked dermatomes and maximal postoperative pain scores (P = .06) or between the number of blocked dermatomes and the PDR on the TPVB side (P = .15).

CONCLUSIONS: This proof-of-concept trial suggests that the effect of TPVB could be monitored by measuring the PDR after anterior chest wall stimulation in the dermatome of interest.

Published ahead of print May 4, 2017.

From the *Institut Curie, Hôpital René Huguenin, Saint-Cloud, France; and Institut Gustave Roussy, Villejuif, France.

Accepted for publication January 26, 2017.

Published ahead of print May 4, 2017.

Funding: This work was only supported by local institutional funds.

The authors declare no conflicts of interest.

Reprints will not be available from the authors.

Address correspondence to Baptiste Duceau, MD, Department of Anesthesiology, Institut Curie-Hôpital René Huguenin, 35, rue Dailly 92210 Saint-Cloud, France. Address e-mail to

Breast cancer remains the most common neoplasia in women and surgery is often part of therapy. Although general anesthesia is still the standard technique for women undergoing major breast surgery, thoracic paravertebral block (TPVB) is now recommended to reduce postoperative pain, nausea, and vomiting.1,2

The thoracic paravertebral space is a small space located next to the vertebral column where the spinal nerves emerge from the vertebral foramen. Its posterior (deep) and anterior (superficial) boundaries are formed by the parietal pleura and the costotransverse ligament prolonged by the internal intercostal membrane. TPVB can be difficult to perform and there is no gold standard to assess the clinical success of this technique. Common clinical evaluations of sensory block include qualitative tests such as the pinprick or cold sensation, which are based on weak stimuli. Thermographic imaging may objectively predict the extent of TPVB, because skin temperature after sympathetic block seems to be correlated to a loss of the pinprick sensation.3 However, thermographic imaging and the loss of the pinprick or cold sensation are surrogate markers and cannot confirm the block of nociception after regional anesthesia.4 For example, electrical stimulation is frequently painful during epidural anesthesia in the absence of a pinprick or cold sensation.5

Pupillometry is a promising tool to monitor analgesia and is useful in both the operating room6–11 and intensive care units.12,13 The pupil dilates in a distinct manner known as pupillary dilation reflex (PDR)6 in response to painful electrical stimulation if the depth of analgesia is insufficient.

The goal of this study was to evaluate the feasibility of recording the PDR in response to a noxious stimulus under general anesthesia to assess analgesia in an anesthetized thoracic dermatome after TPVB.

Back to Top | Article Outline


Trial Design

This prospective single-center observational study evaluated 34 consecutive patients who received TPVB from November 2015 to February 2016. Patients and evaluators were not blinded. The institutional review board approved the study (Commission Scientifique des Essais Thérapeutiques, Hôpital Gustave Roussy, November 2015) and written informed consent was obtained from all patients. The methodology followed the Equator guidelines for observational studies (TREND statement in the supplementary appendix).

Back to Top | Article Outline

Patient Population

Patients scheduled to undergo major breast surgery, tumorectomy or mastectomy, with or without axillary lymph node dissection (ALND), were prospectively included in the study. Exclusion criteria were patients younger than 18 years of age, the presence of ophthalmologic (recent eye surgery or cataracts) or neurologic diseases (diabetic neuropathy, spinal cord injury, or postherpetic neuralgia) that could interfere with pupillometry, the usual contraindications to local anesthetics, local sepsis, severe coagulopathy, and bilateral breast surgery.

Back to Top | Article Outline


Ultrasound (US)-guided TPVB was performed at the second or third thoracic dermatoma (underneath the second or third thoracic vertebra, respectively) under aseptic conditions in the postanesthesia care unit (PACU) before surgery. The patient was placed in a lateral position with the affected side facing up. Baseline electrocardiogram, heart rate, arterial pressure, and oxygen saturation was recorded, and secure IV access was obtained. US was used throughout the procedure: the 5- to 12-MHz linear array probe (Edge ultrasound system; FUJIFILM SonoSite Inc, Bothell, WA) was in a sterile sheath and placed in a transverse position at the second or third thoracic spine. It was moved up and down to identify the transverse process and pleura, as described by Hara et al.14 The needle was inserted in an in-plane fashion and hydrolocalization was used to identify the paravertebral space. Twenty milliliters of 0.75% ropivacaine was slowly injected after negative aspiration to confirm the absence of air or blood while the opening of the paravertebral space was observed on US by the depression of the pleura.

Back to Top | Article Outline

Clinical Evaluation

The onset of sensory block was assessed by a loss of cold discrimination at least 10 minutes after the injection. Each of the dermatomal segments between T2 and T12 was tested. Persistence of any cold sensation was considered to be an absence of sensory block. The level of pain of the injection was assessed using a numeric rating scale.

Back to Top | Article Outline

General Anesthesia

Standard monitors were applied when the patient arrived in the operating room (noninvasive arterial blood pressure, scope, oxygen saturation, entropy; Aisys Carestation GE Healthcare, Little Chalfont, UK). General anesthesia was standardized: after preoxygenation, general anesthesia was induced by 2.5 mg/kg1 propofol and target-controlled infusion of remifentanil (Injectomat TIVAs Agilia; Fresenius Kabi, Bad Homburg, Germany, with Minto pharmacokinetic model15) with an initial remifentanil effect-site concentration of 5.0 ng/mL1. The patient was either intubated after infusion of 0.6 mg/kg1 atracurium or a laryngeal device was used. Volume-controlled ventilation was set between 6 and 7 mL/kg1. General anesthesia was maintained by sevoflurane or desflurane to maintain entropy between 40 and 60. At the end of the surgery, 1000 mg paracetamol and 50 mg paracetamol were infused in the absence of contraindications. In the PACU, tramadol, nefopam, and morphine infusions were used according to local protocols.

Back to Top | Article Outline


The AlgiScan (IDMED, Marseille, France) system was used to record pupil diameters with no influence of the light reflex. At the end of the surgery, while the patient was still under anesthesia, 2 sets of 2 electrodes were placed on the patient’s anterior chest wall, 1 set on each side of the sternum. Electrodes from each set were placed side to side on a horizontal line in the T4 dermatome (nipple line). Two sides of the same dermatomal segment were tested. One side was anesthetized by TPVB and the other side was used as a control. Thus, each patient served as her own control. The same investigator recorded the PDR in response to standardized 5-second 100-Hz/60-mA tetanic stimulation delivered sequentially in the anesthetized (TPVB side) and control dermatome. The PDR was recorded in the eye opposite the block, and close attention was paid to contralateral eye closure. The PDR was the maximal increase in the pupil diameter after a noxious stimulus, expressed as a percentage of initial pupil diameter. Remifentanil was stopped at least 5 minutes before the measurements to obtain nonconfounding effect-site concentrations of remifentanil during PDR records.

Back to Top | Article Outline


After the recovery phase of anesthesia, patients were transferred to the PACU for a 2-hour observation period. Pain was evaluated every 30 minutes at rest and during shoulder mobilization by a nurse who was blinded to the results of the pupillary responses. A numeric rating scale was used, ranging from no pain (0) to unbearable pain (10).

Back to Top | Article Outline

End Point

The primary outcome was the difference between PDR after stimulation on the TPVB side and PDR after stimulation on the control side.

Back to Top | Article Outline

Statistical Analysis

Results are reported as medians (interquartile range [IQR]) or means (standard deviation) for continuous variables and numbers (%) for categorical variables. Because the Kolmogorov-Smirnov test rejected the hypothesis of a normal distribution of the difference in paired PDR, the Wilcoxon signed-rank test for paired samples was used to compare the PDR between the TPVB and control sides. Data for the primary outcome are presented as point estimates of the difference in medians between paired observations, with 95% confidence intervals, obtained by Hodges-Lehmann’s method.16 Linear regressions between the maximal postoperative pain score and the PDR on the TPVB side between the maximal postoperative pain score and the number of dermatomes with a clinical loss of the cold sensation as well as between the number of dermatomes with a clinical loss of the cold sensation and the PDR on the TPVB side were performed. A P value of <.05 was considered to be statistically significant. Software R (version 3.2.2 for Macintosh, licenses GNU GPL; The R Foundation for Statistical Computing, Vienna, Austria) was used to perform all analyses. Based on preliminary data, we determined that 32 patients needed to be included to give the study a power of 90% to detect a 10% point difference with an assumed standard deviation of a 16.7% point difference for the primary outcome at a 2-sided α level of .05 with a Student paired t test. Because conclusions on the primary outcome could not be determined by the Student paired t test, we used the Wilcoxon signed-rank test, a nonparametric test, which could result in a loss of power.

Back to Top | Article Outline


Two of 34 consecutive patients who underwent TPVB could not be included (isolated ALND). Thirty-two American Society of Anesthesiologist I–II–III patients scheduled for major breast surgery were included. Table 1 shows the general characteristics of the study population and the type of surgery. TPVB was performed at the second (94%) or third (6%) thoracic vertebra. Figure 1 shows the extent of the clinical loss of the cold sensation after a mean of 15 minutes (standard deviation 6.5 minutes). A tracheal tube was inserted in 70% and a laryngeal device in 30% of patients. Sevoflurane (62%) or desflurane (38%) was used to provide general anesthesia.

Table 1.

Table 1.

Figure 1.

Figure 1.

At the time PDR were measured, the mean (standard deviation) remifentanil effect-site concentration was 0.06 ng/mL1 (0.09), the expired fraction of sevoflurane was 1.4 (0.7), or the expired fraction of desflurane was 2.5 (1.6).

The median (IQR) PDR were 9% (4%–13%) and 41% (27%–66%) on the TPVB side and on the control side, respectively (Figure 2). There was a significant difference in the PDR between TPVB side and control side with a median difference of 37% points (95% confidence interval [CI], 25–52, P < .001). The initial pupil diameter was 2.3 mm (0.6) without a difference between the initial pupil diameter for the 2 stimulations.

Figure 2.

Figure 2.

Postoperative pain scores (median [IQR]) in the PACU were 1 (0–3) at rest and 1 (0–3) during mobilization, respectively. No patient required morphine titration, 7 patients (22%) received tramadol (100 mg), and 3 patients (9%) received nefopam (20 mg) during the PACU stay.

Figure 3 shows maximal postoperative pain score depending on the PDR recorded on the TPVB side. There was a linear correlation between pain scores and the PDR with a Pearson’s correlation coefficient r = 0.40 (95% CI, 0.06–0.66, P = .02). The difference between the PDR on TPVB side and the PDR on the control side was not correlated to postoperative pain score (P = .89). No correlation was found between the number of blocked dermatomes and the maximal postoperative pain scores (r = −0.35, 95% CI, −0.63 to 0.02, P = .06) or between the number of blocked dermatomes and the PDR recorded on the TPVB side (r = −0.27, 95% CI, −0.58 to 0.10, P = .15).

Figure 3.

Figure 3.

No local complications occurred owing to pupillometry. Nausea, vomiting, and the use of antiemetic therapy during PACU were rare.

Back to Top | Article Outline


This prospective single-center, observational study showed that the PDR was diminished by TPVB performed before general anesthesia in patients scheduled for breast surgery.

Monitoring TPVB by the PDR could improve the quality of care. If the PDR can predict ineffective TPVB before the patient has recovered, the clinician can prevent acute postoperative pain by administering analgesic medications or infiltrating the surgical wounds with local anesthetics.17 On the 1 hand, we believe that the PDR is more interesting than other preoperative predictors such as the loss of cold sensation to predict TPVB failure. Indeed, the PDR can be recorded immediately before the patient recovers consciousness, when TPVB should be the most effective. On the other hand, the clinical loss of cold sensation or the pinprick test can only be assessed before surgery when the onset time and the quality of blockade can result in false-negative results or after recovery when the assessment of blocked dermatomes has no value to adjust the analgesic approach. In this study, loss of the cold sensation was evaluated after a mean of 15 minutes, which is premature because the onset time of ropivacaine is more than 20 minutes.18 Delaying surgery to assess the loss of the pinprick and cold sensations after 40 minutes might have improved the correlation between the sensory test and the pain scores in the PACU as well as between the sensory test and the PDR. Nevertheless, the PDR provides objective results, unlike sensory tests, which depend on patient sensitivity and understanding.

Moreover, if a clinician could reliably control the effectiveness of TPVB, he or she could more easily decrease perioperative opioid infusion. This is highly important because decreasing perioperative opioid use could influence long-term outcomes such as cancer recurrence19 or chronic postoperative pain.20

Initial pupil size by general anesthesia was similar to previous findings.7,21 Numerous drugs modify pupil diameter under general anesthesia. All measurements were made before the use of a reversal agent for neuromuscular blocking. Droperidol was only used as a rescue treatment in the PACU for nausea or vomiting because neuroleptic drugs have a pupilloconstrictive effect.22 Finally, there was no difference between basal pupil size between the 2 measurements, meaning that pupil diameter promptly returned to its baseline after the PDR measurements on both sides. The PDR expressed as percentages of initial pupil diameters can thus be compared in this study.

Our study was powered to detect a 10% point difference based on preliminary data. These data were obtained with a similar protocol but posterior chest wall stimulation. The difference found in the present study with anterior chest wall stimulation was more robust. Thus, although the present study was overpowered, the results are clinically relevant with a median difference of 37% points.

The correlation between pain scores and the PDR on the TPVB side was weaker than expected. More than half of the patients had ALND in addition to their breast surgery and felt pain in the axilla and inner arm. These areas can be under brachial plexus innervation, which is not anesthetized by TPVB, or a T1 intercostal nerve, which could be missed by insufficient cephalic spread of TPVB. Our assessment of overall pain without discriminating between locations probably led to a confusion factor explaining the weak relationship between pain scores and the PDR. Moreover, deep planes of the chest are under medial pectoral and lateral pectoral nerves (innervation of pectoralis major and pectoralis minor muscles), which also derive from brachial plexus.

Our study has certain limitations. The site of noxious stimulation in the patient’s anterior chest wall creates 3 issues: first, the manufacturer of Algiscan recommends stimulation of the ulnar nerve at the wrist; thus, it has not been validated in other areas. However, pupillometry has been successfully performed in the ankle to evaluate popliteal sciatic nerve block8 and with C5 and L4 dermatome stimulation during epidural anesthesia.23 Larson et al6 observed that the PDR accurately predicted the level of sensory block within 2 dermatomes during combined epidural–general anesthesia. A neutral stimulation site (such as the ulnar nerve) should have been used for validation of our data. Second, pupil constriction by opiate drugs is well known and has been described with remifentanil.21 Opioid-induced pupillary constriction could bias the PDR. Remifentanil, a short-acting opioid, was stopped at least 5 minutes before the PDR was recorded, resulting in a low and probably nonconfounding effect-site concentration of remifentanil.21 Third, the common practice of long-acting opioid administration probably limits the generalizability of this strategy. Nonetheless, PDR shows good results for monitoring the analgesia level with long-acting opioid like alfentanil9 during anesthesia and sufentanil10 in the immediate postoperative period or fentanyl13 in intensive care. The validation of our strategy with long-acting opioids still needs further studies.

TPVB is known to have a low side effect rate,24 and there were no side effects from TPVB in this small cohort.

In conclusion, this feasibility study assessed and confirmed the value of PDR in detecting the success of unilateral TPVB during general anesthesia. A marked difference was found between the PDR on the blocked side and the unblocked side. This approach supports previous studies using the PDR to evaluate sensory blockade during general anesthesia, but further studies comparing the PDR with sensory tests are needed to show improvement in the quality of care with this approach.

Back to Top | Article Outline


Thanks to Mehdi Benichou (London, England) and Ugo Coltel (Cork, Ireland) who critically reviewed the proposal and Dale Roche-Lebrec (Paris, France) for language editing.

Back to Top | Article Outline


Name: Baptiste Duceau, MD.

Contribution: This author helped in study design, patient recruitment, data collection and analysis, and writing up of the first draft of the paper.

Name: Mélanie Baubillier, MD.

Contribution: This author helped in study design and patient recruitment, data collection, and analysis.

Name: Gaëlle Bouroche, MD.

Contribution: This author helped in study design and critical revising of the draft.

Name: Aline Albi-Feldzer, MD.

Contribution: This author helped in study design, patient recruitment, and critical revising of the draft.

Name: Christian Jayr, MD, PhD.

Contribution: This author helped in study design, patient recruitment, critical revising, and final approval of the draft.

This manuscript was handled by: Richard Brull, MD, FRCPC.

Back to Top | Article Outline


1. Comité douleur-anesthésie locorégionale et le comité des référentiels de la Sfar. [Formalized recommendations of experts 2008. Management of postoperative pain in adults and children] [in French]. Ann Fr Anesth Réanim. 2008;27:1035–1041.
2. 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;105:842–852.
3. Cheema SP, Ilsley D, Richardson J, Sabanathan S. A thermographic study of paravertebral analgesia. Anaesthesia. 1995;50:118–121.
4. Curatolo M, Petersen-Felix S, Arendt-Nielsen L. Sensory assessment of regional analgesia in humans: a review of methods and applications. Anesthesiology. 2000;93:1517–1530.
5. Curatolo M, Kaufmann R, Petersen-Felix S, Arendt-Nielsen L, Scaramozzino P, Zbinden AM. Block of pinprick and cold sensation poorly correlate with relief of postoperative pain during epidural analgesia. Clin J Pain. 1999;15:6–12.
6. Larson MD, Sessler DI, Ozaki M, McGuire J, Schroeder M. Pupillary assessment of sensory block level during combined epidural/general anesthesia. Anesthesiology. 1993;79:42–48.
7. Huybrechts I, Barvais L, Ducart A, Engelman E, Schmartz D, Koch M. Assessment of thoracic epidural analgesia during general anesthesia using pupillary reflex dilation: a preliminary study. J Cardiothorac Vasc Anesth.2006;20:664–667.
8. Isnardon S, Vinclair M, Genty C, Hebrard A, Albaladejo P, Payen JF. Pupillometry to detect pain response during general anaesthesia following unilateral popliteal sciatic nerve block: a prospective, observational study. Eur J Anaesthesiol. 2013;30:429–434.
9. Constant I, Nghe MC, Boudet L, et al. Reflex pupillary dilatation in response to skin incision and alfentanil in children anaesthetized with sevoflurane: a more sensitive measure of noxious stimulation than the commonly used variables. Br J Anaesth. 2006;96:614–619.
10. Aissou M, Snauwaert A, Dupuis C, Atchabahian A, Aubrun F, Beaussier M. Objective assessment of the immediate postoperative analgesia using pupillary reflex measurement: a prospective and observational study. Anesthesiology.2012;116:1006–1012.
11. Girard M-A. La pupillométrie comme mesure d’analgésie d’une technique d’ALR sous anesthésie générale. [Pupillometry to evaluate peripheral verve block under general anesthesia]. Paper presented at: Congress of the French Society of Anesthesia and Intensive Care Medicine; 2011.
12. Payen JF, Isnardon S, Lavolaine J, Bouzat P, Vinclair M, Francony G. Pupillometry in anesthesia and critical care. Ann Fr Anesth Reanim.2012;31:e155–e159.
13. Paulus J, Roquilly A, Beloeil H, Théraud J, Asehnoune K, Lejus C. Pupillary reflex measurement predicts insufficient analgesia before endotracheal suctioning in critically ill patients. Crit Care.2013;17:R161.
14. Hara K, Sakura S, Nomura T, Saito Y. Ultrasound guided thoracic paravertebral block in breast surgery. Anaesthesia.2009;64:223–225.
15. Minto CF, Schnider TW, Egan TD, et al. Influence of age and gender on the pharmacokinetics and pharmacodynamics of remifentanil. I. Model development. Anesthesiology.1997;86:10–23.
16. Lehmann EL. Nonparametric confidence intervals for a shift parameter. Ann Math Stat. 1963;34:1507–1512.
17. Albi-Feldzer A, Mouret-Fourme EE, Hamouda S, et al. A double-blind randomized trial of wound and intercostal space infiltration with ropivacaine during breast cancer surgery: effects on chronic postoperative pain. Anesthesiology.2013;118:318–326.
18. Gaudreault F, Drolet P, Fallaha M, Varin F. Modeling the anesthetic effect of ropivacaine after a femoral nerve block in orthopedic patients: a population pharmacokinetic–pharmacodynamic analysis. Anesthesiology.2015;122:1010–1020.
19. Sessler DI, Ben-Eliyahu S, Mascha EJ, Parat MO, Buggy DJ. Can regional analgesia reduce the risk of recurrence after breast cancer? Methodology of a multicenter randomized trial. Contemp Clin Trials.2008;29:517–526.
20. Andreae MH, Andreae DA. Regional anaesthesia to prevent chronic pain after surgery: a Cochrane systematic review and meta-analysis. Br J Anaesth. 2013;111:711–720.
21. Barvais L, Engelman E, Eba JM, Coussaert E, Cantraine F, Kenny GN. Effect site concentrations of remifentanil and pupil response to noxious stimulation. Br J Anaesth. 2003;91:347–352.
22. Larson MD. The effect of antiemetics on pupillary reflex dilation during epidural/general anesthesia. Anesth Analg. 2003;97:1652–1656.
23. Larson MD, Berry PD, May J, Bjorksten A, Sessler DI. Autonomic effects of epidural and intravenous fentanyl. Br J Anaesth.2007;98:263–269.
24. Davies RG, Myles PS, Graham JM. A comparison of the analgesic efficacy and side-effects of paravertebral vs epidural blockade for thoracotomy–a systematic review and meta-analysis of randomized trials. Br J Anaesth.2006;96:418–426.
Copyright © 2017 International Anesthesia Research Society