Advanced therapy in thoracic surgery - part 1 pot

The main blocks used for thoracotomy patients are intercostal nerve blocks, interpleural analgesia, thoracic paravertebral nerve blocks TPVBs, and epidural analge-sia.. For example, Kolv

It is natural to want to relieve pain and suffering None are

more aware of this than those professionals who have

devoted their lives to the provision of anesthesia, yet we

have often been prevented from alleviating pain by not

understanding its pathogenesis or by a lack of appropriate

tools to deal with it Intraoperative pain is now only of

historic concern It is our fervent hope that postoperative

pain will follow intraoperative pain into the history books

Not so long ago, certainly within the professional

experience of some of us, a minimalist approach was

taken to the management of pain after thoracic surgery

Anesthesiology residents and faculty alike were

admon-ished to keep total opioid dosage low so the patient

would “want to breathe” after surgery During this era,

the classic thoracotomy patient would be nearly apneic

from pain in the postanesthesia care unit Hypoxic and

hypercarbic, diaphoretic and hypertensive, patients

would gradually improve to the point at which they

could actually breathe and complain of pain only after

large doses of opioids Frequent arterial blood gas

analy-ses often demonstrated the unusual observation that the

administration of opioids led to a decrease in carbon

dioxide tension and an increase in oxygen tension in this

setting

In 1973 Gibbons and colleagues suggested that

thoracic epidural blockade was the treatment of choice

for relief of pain after a chest injury.1The major limitation

was sympathetic blockade causing hypotension To

prevent this complication, they advocated intercostal

blockade for fractures at or above the fifth rib The

modern era of pain management after thoracic surgery

began with the introduction of epidural narcotic

tech-niques for post-thoracotomy pain.2,3Soon continuousinfusions were advocated,4and the effect of better postop-erative analgesia on pulmonary function was investigated.5

This led to an increased ability to control tomy pain and also stimulated the overall interest in find-ing other useful modalities for post-thoracotomy painrelief Older or abandoned techniques were investigatedwith renewed interest and used singly or in combinations.For at least the past 10 years, the immediate postopera-tive pain of most thoracotomy patients has been wellhandled There are occasional patients whose pain is diffi-cult to manage because of coexisting disease processesthat contraindicate epidural analgesia, anatomic factors,and/or pre-existing chronic pain, but currently there aretechniques to help even these patients As an uninten-tional consequence of relieving the severe, acute incisionalpain of surgery, we may have unmasked other sources ofequally troubling pain such as referred pain and sympa-thetically mediated pain Much research is focused ontreating these “new” modalities This unbundling of post-

post-thoraco-operative pain has been termed disaggregation.6Anotherarea of increasing interest is the pathogenesis of chronicpostoperative pain Whether we can affect or even preventthis unhappy outcome remains to be seen

Post-Thoracotomy Pain

Acute Pain

Pain in the first few weeks after a thoracotomy arisesfrom a variety of different mechanisms The best charac-terized mechanism is somatic pain, which is localized tothe area around the incision and chest tube insertion

sites It is produced by direct injury to the skin and

underlying subcutaneous tissues, fasciae, ligaments,

muscles, and ribs Damaged tissue releases a variety of

algesic substances, including substance P, prostaglandins,

and serotonin, which stimulate the peripheral nerve

endings.7Intercostal nerves from the area conduct these

pain impulses to the spinal cord and thence to the brain

via the spinothalamic and spinoreticular tracts Somatic

pain is responsible for the sharp, severe postoperative

pain that is exacerbated by movement and is believed to

be primarily mediated by type A delta nerve fibers.8

Visceral, or nonincisional, pain is responsible for the dull,

nauseating, diffuse thoracic wall “aching” sensation

expe-rienced after a thoracotomy It is mediated by type C

nerve fibers, which travel with the autonomic nerves

Both the vagus and sympathetic ner ves probably

contribute to this type of pain.8

Another form of pain frequently reported in

post-thoracotomy patients is localized to the ipsilateral

shoul-der region Although it is often moshoul-derate to severe in

intensity and present in 75 to 85% of patients who have

had a thoracotomy,9–11this type of pain has received little

attention in the literature It has been attributed to a

vari-ety of factors, including distraction of the posterior

thoracic ligaments or shoulder joint due to patient

posi-tioning; stretching of the brachial plexus, also as a

conse-quence of intraoperative positioning; transection of a

major bronchus; and referred pain from the phrenic

nerve.9As the latter provides sensory innervation to the

pericardium and pleura, mechanical trauma to these

regions during surgery and irritation of the pleural

surfaces by chest tubes postoperatively can result in

phrenic nerve stimulation, with referral to the shoulder

Scawn and colleagues have demonstrated a reduction in

the incidence of post-thoracotomy shoulder pain from

85 to 33% with the injection of 10 mL of 1% lidocaine

into the periphrenic fat at the level of the diaphragm.9In

this same study, there was a small but insignificant

increase in arterial partial pressure of carbon dioxide

(PaCO2) in the first 2 postoperative hours in patients

receiving a phrenic nerve block, thereby suggesting the

possibility of diaphragmatic paresis The technique may

thus be inappropriate in patients with severely

compro-mised respiratory function The lack of efficacy of

supras-capular nerve blockade in relieving post-thoracotomy

shoulder pain demonstrated by Tan and colleagues

provides further evidence that distraction of the shoulder

joint does not play a major role in the generation of this

type of pain.11

The extent to which the surgical approach contributes

to the severity of post-thoracotomy pain is unclear

Anteroaxillary and anterior limited thoracotomies are

less painful procedures than are posterolateral

thoraco-tomies.12,13 When muscle-sparing thoracotomies havebeen compared with traditional posterolateral thoraco-tomies (involving a transection of the latissimus dorsimuscle), some studies have demonstrated less postopera-tive pain with the former,14whereas others have revealed

no difference between the two techniques.15,16It is wellappreciated that thoracoscopic procedures result in lesspain than do traditional thoracotomies in the early post-operative period, but Nomori and colleagues havedemonstrated this benefit to be lost by 14 days aftersurgery.17The lack of a consistent and/or persistentdecrease in post-thoracotomy pain with less extensivesurgical incisions provides further evidence that theactual surgical incision is just one of several mechanismsresponsible for post-thoracotomy pain

Chronic Pain

Post-thoracotomy pain syndrome is defined as “pain thatrecurs or persists along a thoracotomy scar at least twomonths following the surgical procedure.”18 There is

“usually tenderness, sensory loss, and absence of sweatingalong the thoracotomy scar.”18The incidence is variable,ranging from 2 to 67%.19Dajczman and colleagues stud-ied 59 of 206 sequential patients who had undergone aunilateral thoracotomy; all procedures were performed

by one surgeon over a period of 5 years.20Thirty to 73%

of the patients available for evaluation were experiencingpain (Table 1-1), which most rated at a visual analogscale (VAS) of two to four (Figure 1-1)

These results were confirmed by Perttunen andcolleagues,21who found an incidence of post-thoracotomypain of 80% at 3 months, 75% at 6 months, and 61% after

1 year More than 50% of these patients had limitations oftheir activities of daily living imposed by the chronic pain.There was also a 3 to 5% incidence of severe pain.Intriguingly, early consumption of larger quantities ofnonsteroidal anti-inflammatory drugs (NSAIDs) wasassociated with an increased incidence of long-term prob-lems As suggested by Perttunen and colleagues, this couldTABLE 1-1 Frequency of Post-Thoracotomy Pain at Various Intervals following Surgery

Adapted from Dajczman E et al 20

*At least 2 months post-thoracotomy.

influence of the variable permeability of the skin is

decreased Nevertheless, there is a significant variability

in the systemic drug levels and analgesic effects As well,

there is an accumulation of fentanyl under the patch,

providing appreciable serum levels for up to 24 hours

after patch removal.25Because of the possibility of apnea

owing to high serum levels in opioid-naive patients,26

transdermal fentanyl is not currently recommended for

acute postoperative pain One approach that may permit

its use in the future is to combine a low-dose

transder-mal fentanyl patch with an NSAID.27Another possibility

is to add electrical control to enhance the rate of

fentanyl absorption across the skin This iontophoretic

route of administration is currently experimental but

may offer the possibility of patient-controlled

transder-mal fentanyl in the future.28

Fentanyl can be delivered across the mucous

mem-branes of the mouth Oral transmucosal fentanyl citrate

(OTFC, Actiq Abbott Laboratories, Abbott Park, IL) has

been used for breakthrough chronic cancer pain as well

as acute postoperative pain.29,30It would be a good choice

if the intravenous route was temporarily unavailable for

acute postoperative analgesia Fentanyl, sufentanil,

butor-phanol, heroin, oxycodone, and meperidine have been

administered through the nasal mucosa, and morphine,

codeine, fentanyl, heroin, and hydromorphone have been

administered by inhalation.31

Ketamine

Ketamine is a phencyclidine derivative occasionally

employed as an anesthetic induction agent Uniquely

among induction agents, it produces what has been

termed dissociative anesthesia The analgesia does outlast

the anesthetic effects and occurs at lower serum levels, so it

can be useful in a decreased dosage as a postoperative

anal-gesic Ketamine may be given by intravenous,

subcuta-neous, epidural (see below), oral, and transdermal routes.32

Ketamine produces analgesia by multiple

mecha-nisms, including inhibition of N-methyl-D-aspartate

(NMDA) receptors, depression of the thalamus while

activating the limbic system, and direct spinal effects

NMDA receptors are involved in hyperalgesia or

neuro-pathic pain, which suggests that ketamine would be a

good choice for analgesia for these patients.33A recent

study in rats demonstrated that ketamine had different

mechanisms of action depending on the presence or

absence of inflammation Antinociceptive effects were

created by activation of the monoaminergic descending

inhibitory system, whereas in a hyperalgesic state

induced by inflammation, inhibition of NMDA

activa-tion was the likely mechanism of the antihyperalgesia.34

Ketamine is a useful agent when narcotics and

neurax-ial agents are contraindicated or working poorly Chow

and colleagues described a patient undergoing multiplethoracotomies whose pain management was complicated

by infection and the development of neuropathic pain.35

Low-dose ketamine was used to decrease the need fornarcotics after his fourth thoracotomy, with good results

It has also been suggested that ketamine should have emptive effects because of its action at NMDA receptors

pre-A landmark study in cholecystectomy patients found lesspostoperative pain, as measured by VAS scores andmorphine consumption, in the group given low-doseintraoperative ketamine.36An alternative explanation forthe observed improved analgesia is that ketamineprevents the development of acute tolerance to opioids.37

Nonsteroidal Anti-inflammatory Drugs

NSAIDs have proven to be a useful component of erative pain relief Many oral NSAIDs have been usedincluding ibuprofen, naproxen, and ketoprofen The onlycurrently available parenteral NSAID is ketorolactromethamine (Toradol, Roche Laboratories, Nutley, NJ).The addition of ketorolac to a patient-controlled epiduralanalgesia (PCEA) regimen employing hydromorphonealone significantly decreased the incidence of noninci-sional pain.38Ketorolac is also employed in the treatment

postop-of breakthrough pain with otherwise satisfactory epiduralanalgesia Ketorolac has several other features that make ituseful in postoperative thoracotomy patients Theseinclude its moderate potency (equivalent to morphine insome studies39); ease of administration by the intravenousand intramuscular routes; lack of acute tolerance, whichmay occur with even a single dose of opioid40; and lack ofsignificant cardiorespiratory or central nervous systemside effects

NSAIDs inhibit cyclooxygenase (COX), the enzymethat regulates the conversion of arachidonic acid toprostaglandins There are two isoenzyme forms of COX.COX-1 is always present (constitutive) It modulatesplatelet activity and gastrointestinal cytoprotection and

is involved in maintaining renal function in volemic states COX-2 is thought to be inducible byinflammatory stimuli and is involved with inflamma-tion and pain Conventional NSAIDs, such asindomethacin and ketorolac, which inhibit both COX-1and COX-2, have been implicated in postoperativebleeding and gastric ulceration They also may predis-pose to renal failure if the patient is concomitantly hypo-volemic or even just relatively “dry,” as post-thoracotomypatients often are Specific inhibitors of COX-2 weredeveloped in an attempt to prevent the side effects ofconventional NSAIDs while maintaining the benefits.Current selective COX-2 inhibitors still exhibit somepredilection for causing renal failure and gastric ulcera-tion but debatably to a lesser extent than conventional,

hypo-nonselective NSAIDs.41–43 The selective COX-2 inhibitors

do not affect platelet function and have not been shown

to increase postoperative blood loss As a result they can

be used perioperatively with relative impunity from

hemorrhage As of this writing, there is no parenteral

COX-2 inhibitor available, although one is in US Food

and Drug Administration trials

Regional Analgesia Techniques

In the past two decades, regional analgesia techniques

have become the primary means of providing optimal

pain relief after a thoracotomy Although the type C

nerve fibers responsible for autonomically mediated

visceral pain have abundant opioid receptors, type A

delta nerve fibers, which mediate somatic incisional pain,

contain a paucity of these receptors.4 4Accordingly,

systemically administered opioids have limited efficacy in

controlling acute post-thoracotomy pain, especially that

associated with activity In contrast, local anesthetics,

which are an integral component of most regional

anal-gesia techniques, are very effective in blocking

conduc-tion in both type A delta and C nerve fibers

The main blocks used for thoracotomy patients are

intercostal nerve blocks, interpleural analgesia, thoracic

paravertebral nerve blocks (TPVBs), and epidural

analge-sia Characteristics of these blocks are summarized in

Table 1-5 Each may be performed as a single injection

through a needle, but owing to the prolonged period of

substantial pain experienced after a thoracotomy, catheter

techniques are used more commonly (with the possible

exception of intercostal nerve blocks, as discussed below)

A standard 18- or 20-gauge epidural catheter may be

placed through a hollow needle into the appropriate area

for each block, and analgesic medication is administered

through this catheter either as intermittent boluses or acontinuous infusion The former has the disadvantage ofsupplying fluctuating levels of analgesic in the area of theblock and thus providing varying degrees of pain relief forthe patient The latter has the disadvantages of providingmore analgesic than is necessary during periods of lesspainful stimulation, and promoting the accumulation ofanalgesic medication over time,45unless appropriatedecrements in infusion rates are made With the goal ofminimizing the disadvantages of both methods, lowcontinuous (“basal”) infusion rates have been combinedwith intermittent boluses administered on an as-neededbasis (which is usually patient controlled)

Regional analgesia use in thoracotomies has severalunique features compared with use in other types ofsurgery First, all techniques except epidurals may beperformed under direct vision from an internal approachbefore the chest is closed This not only increases the easewith which the blocks are performed, but may also improvetheir success rate when compared with blocks performedvia percutaneous techniques (although no studies havedirectly addressed this issue) As well, the risk of developing

a pneumothorax, which is a potentially limiting factor forintercostal nerve blocks and interpleural analgesia, is irrele-vant because the thoracic cavity is open intraoperativelyand chest tubes are used postoperatively Finally, hypov-olemia is a relatively common occurrence in patients afterthoracotomy because extensive fluid administration hasbeen implicated in the development of postoperativepulmonary edema, especially after pneumonectomy.46

Therefore, regional analgesia techniques producing sive blockade of the sympathetic nervous system andperipheral vasodilation may be accompanied by a signifi-cant risk of hypotension, and are often avoided

exten-TABLE 1-5 Summary of Factors Related to Regional Analgesia Techniques*

Technique Ease of Analgesic Preservation Modification Hypotension Motor Blockade Urinary Retention Respiratory

= not a factor; ± = sometimes a factor; + to ++++ = degrees of being somewhat a factor to being an important factor.

*For post-thoracotomy pain.

† With opioid/low-dose local anesthetic infusions.

Of the four types of regional analgesia discussed in

this chapter, epidural analgesia is the only technique for

which agents other than local anesthetics have been

successfully used This is not surprising because the

intercostal nerve block, interpleural analgesia, and

paravertebral nerve block techniques depend primarily

on blocking impulse transmission within somatic nerves

By contrast, blockade of pain pathways within the spinal

cord may be accomplished by other drugs, most

commonly opioids, delivered into the epidural space

Although several local anesthetic agents are available,

bupivacaine has been the most popular choice for

post-thoracotomy blocks over the past couple of decades,

primarily because of its prolonged duration of action

Concentrations of 0.25 to 0.5% are necessary to provide

adequate sensory blockade with most of the blocks

discussed below, although lower concentrations have been

used in the epidural space, when combined with opioids

Since its release in 1996, ropivacaine has been used

increasingly for a variety of intraoperative and

postopera-tive situations, and although the current literature

regard-ing post-thoracotomy regional analgesia focuses on

bupivacaine, ropivacaine will probably play a major role

in clinical practice and the literature in the future It is an

amide local anesthetic structurally similar to bupivacaine

that has the unique quality of being supplied as the pure

S-(
)-enantiomer This contrasts with the other local

anesthetics, which exist as racemic mixtures of both the

ropiva-caine produces less cardiovascular and central nervous

system toxicity,47,48similar analgesia, and a less intense and

shorter duration of motor blockade than does

bupiva-caine when administered into the epidural space.49,50

Low concentrations of epinephrine

(1:100,000–1:400,000) are frequently added to the

solu-tion used for the regional analgesia techniques to decrease

the quantity of medication absorbed into the systemic

circulation This should extend the duration and possibly

improve the degree of analgesia, and decrease the risk of

systemic toxicity from the drug Lower peak plasma

concentrations have been convincingly demonstrated

when epinephrine has been added to the solutions used in

intercostal nerve blocks,51interpleural analgesia,52and

epidural analgesia,53,54but data regarding the duration and

quality of analgesia and systemic toxicity are more

vari-able There is even evidence that the addition of

epineph-rine to epidural opioid solutions may increase the

incidence of some opioid-related side effects, especially

pruritus.55–57 Epinephrine may also directly contribute to

the pain relief achieved with epidural analgesic techniques

by stimulation of 2-adrenergic receptors in the dorsal

horn of the spinal cord.58

Ultralong-acting local anesthetics and opioids rently under development have been advocated as ameans of providing prolonged analgesia from a singledose However, their eventual role in the management ofacute post-thoracotomy pain is unclear because prolon-gation of analgesic effects is accompanied by a prolonga-tion of the duration of adverse events, which hasparticular relevance in the case of life-threatening cardio-vascular and respiratory depression

cur-Despite their apparent usefulness in post-thoracotomypatients, regional analgesia techniques are not appropriatefor all individuals Absolute contraindications for all types

of regional analgesia include patient refusal, an allergy tothe medication to be used, a lack of resuscitative equip-ment, a lack of ability to use the resuscitative equipment,and an infection or tumor at the site of injection Relativecontraindications are often specific for the type of blockand are discussed below for the individual techniques.Knowledge of the contraindications may be critical inchoosing the specific block for a particular patient

Intercostal Nerve Block

definition and techniqueIntercostal nerve block is a technique in which a localanesthetic is injected into the immediate vicinity of theintercostal nerve as it lies in the costal groove on theinternal surface of the rib In this position, the intercostalnerve traverses between the internal intercostal and inter-costalis intimus muscles and is located just caudad to theintercostal artery and vein Local anesthetic is injected 7

to 8 cm from the posterior midline, proximal to theorigin of the lateral cutaneous branch in the midaxillaryline.59Because there is a considerable overlap of sensoryinnervation of the thoracic dermatomes, it is necessary toblock at least one level above and below the desireddermatomal level

Intercostal nerve blocks are often performed by a

“single-shot” injection through a needle There is limitedspread of local anesthetic from one intercostal space tothe next; therefore, separate injections at each level areusually necessary For posterolateral thoracotomy inci-sions, intercostal nerve blocks are usually performed atT3 to T7 Three to 5 mL of local anesthetic is adminis-tered with each block; thus, a total of 20 to 25 mL of localanesthetic is used Analgesia persists for 5 to 12 hoursafter a single injection,60–63 and intercostal nerve blocksmay be repeated as necessary A variety of catheter tech-niques have also been described,62,64,65although most ofthese studies involved the use of more than one catheter,which creates a cumbersome situation

mechanism of actionIntercostal nerve blocks produce analgesia by direct

blockade of the intercostal nerves There is usually

mini-mal or no spread of anesthetic proximini-mally to the dorsal

rami of the intercostal nerves or the sympathetic chain

efficacyIntercostal nerve blocks are moderately effective for

post-thoracotomy pain For example, Kolvenbach and

colleagues detected “adequate” analgesia in

approxi-mately 76% of their group of patients, as measured by

the lack of need for supplemental opioids.6 2 When

compared with placebo or parenteral opioids, intercostal

nerve blocks have usually been shown to produce better

pain control with lower pain scores and/or fewer

supple-mental opioids.64–68

Only two studies have compared intercostal nerve

blocks with other regional techniques for post-thoracotomy

pain Asantila and colleagues compared intercostal nerve

blocks with epidural analgesia with either bupivacaine or

morphine, and found no significant differences between

treatments with respect to pain scores or supplemental

parenteral opioid requirements.69More recently, Perttunen

and colleagues randomized 45 patients to receive intercostal

nerve blocks (performed at T3–T7 via an internal approach

and administered as a single injection just prior to wound

closure), TPVBs, or continuous epidural analgesia with

bupivacaine.70In the first 4 hours after surgery, pain scores

during coughing were significantly lower in the intercostal

nerve block group than in the other two groups No

differ-ences were noted in supplemental morphine consumption,

pain scores at rest, or pain scores with coughing after the

initial 4-hour period However, the authors emphasize that

pain relief in all patients was only fair (VAS pain scores of

28–62/100 at rest and 62–91/100 with coughing), and

opti-mizing the management of these techniques may have

produced different results

Analgesic efficacy may be limited in intercostal nerve

blocks owing to a lack of blockade of the dorsal rami,

which can result in persistent pain at the medial edge of

the incision, and muscles and ligaments in the

surround-ing area Failure to block the sympathetic chain, vagus,

and phrenic nerves may further limit the ability of

inter-costal nerve blocks to provide optimal pain relief after

thoracotomy

Intercostal nerve blocks also appear to be moderately

effective in improving pulmonary function This is

suggested in several,64,66,71,72but not all,65studies by higher

values of forced expiratory volume in 1 second (FEV1),

forced vital capacity (FVC), and/or peak expiratory flow

rate (PEFR) in patients receiving these blocks compared

with values in patients receiving a placebo or parenteralopioids Compiling the results of several studies,Richardson and colleagues demonstrated an overall 55%preservation of spirometric function (vs preoperativevalues) with intercostal nerve blocks by 48 hours post-thoracotomy.8Despite the above observations, most stud-ies have failed to demonstrate that intercostal nerveblocks decrease the incidence of postoperative complica-tions in post-thoracotomy patients Furthermore,although Deneuville and colleagues showed that inter-costal nerve blocks were associated with fewer postopera-tive respiratory complications than was as-neededparenteral opioid, the incidence of complications withintercostal nerve blocks was identical to that with “fixed-schedule” intramuscular opioid injections.65

advantages and disadvantagesThe main advantage of intercostal nerve blocks is the easewith which they can be performed.61 They require littletraining and no special equipment The technique is quitesafe, and any significant complication usually occurswithin 30 minutes of performing the block As such, nospecial monitoring is necessary for patients with theseblocks beyond the immediate post-block time period.The main disadvantages of intercostal nerve blocks arethe necessity of performing separate blocks at multiplelevels, and the relatively short duration of analgesiaachieved via the single-injection techniques

adverse effectsThe most common adverse effect associated with the use

of intercostal nerve blocks for thoracotomy is the opment of high systemic blood levels of local anesthetic.This is a consequence of both the volume needed forinjections at multiple levels and the vascularity of thearea of injection Peak blood levels of local anestheticoccur at 5 to 20 minutes,61,64,73and they are higher thanwith interpleural analgesia, TPVBs, and epidural analge-sia.70,74Case reports of spinal anesthesia associated withthe use of intercostal nerve blocks have also beenreported.75,76This has been postulated to be due to retro-grade intraneural spread of local anesthetic to thesubarachnoid space Most cases have involved intercostalnerve blocks performed by an internal approach duringthoracotomy, possibly because of the more medial injec-tion of the local anesthetic in these circumstances

devel-contraindicationsThere are no absolute contraindications specific to inter-costal nerve blocks The main relative contraindication ofintercostal nerve blocks when used for post-thoracotomyanalgesia is in patients for whom the effects of high

systemic blood levels of local anesthetic may be

particu-larly detrimental, which includes patients with cardiac

conduction defects and seizure disorders

Interpleural Analgesia

definition and technique

The term interpleural analgesia refers to a technique

whereby local anesthetic is placed into the interpleural

space, located between the visceral and parietal pleurae

The term intrapleural analgesia is often used

interchange-ably with interpleural analgesia, but the former is

anatomically incorrect For thoracotomy patients, a

multiorifice epidural catheter is usually inserted into the

interpleural space under direct vision by the surgeon

prior to chest closure, and a local anesthetic is

adminis-tered either as a continuous infusion or intermittent

bolus doses Some authors emphasize suturing the

inter-nal tip of the catheter high in the interpleural space (in

the cranial portion of the thoracic cage) to prevent

dislodgment,77whereas others recommend placing the tip

at the level of the incision.78Table 1-6 presents examples

of dosage regimens None has been demonstrated to be

superior to the others

mechanism of actionInterpleural analgesia produces pain relief primarily by

diffusion or bulk flow of local anesthetic through the

parietal pleura, into the subpleural space, and finally to

the intercostal nerves The resultant effect is a multilevel

intercostal nerve block.79Interpleural analgesia

tech-niques may also block other nervous structures including

the vagus and phrenic nerves as they traverse through the

interpleural space,77pain receptors in the parietal pleura,

and the thoracic sympathetic chain, by diffusion of local

anesthetic into the paravertebral space The clinical

importance of blockade at these secondary sites is

unclear and may contribute to the variable results in

studies examining the efficacy of interpleural analgesia

efficacyThe efficacy of interpleural analgesia for post-thoracotomy

pain is controversial.8 0 Compared with placebo or

parenteral opioids, interpleural analgesia has been shown

to improve analgesia in some studies,81,82and to have

minimal or no effect in others.83–85Interpleural analgesia

has also been demonstrated to produce a degree of

anal-gesia similar to TPVB and thoracic epidural analanal-gesia

with bupivacaine in some studies,7 8 , 8 6 but less than

thoracic epidural bupivacaine, lumbar epidural

hydro-morphone, and lumbar epidural morphine in others.74,87,88

The lack of consistent efficacy for post-thoracotomy

pain has been primarily attributed to the loss of local

anesthetic by drainage through chest tubes Ferrante andcolleagues documented a 30 to 40% loss of an injecteddose of bupivacaine over a 4-hour period through thechest tubes.89For interpleural analgesia administered viathe bolus method, clamping the chest tubes for 15 to 30minutes after each dose has been advocated to helpcircumvent this problem,77although the safety and effi-cacy of such a maneuver has been questioned.8

Other factors that may contribute to the lack of gesic efficacy are dilution of local anesthetic with pleuralexudate, and uneven distribution of local anestheticthroughout the pleural space The latter may occurbecause of inflammation of the pleura by the currentsurgical procedure and/or the presence of fibrous tissuefrom previous pleural disease or thoracotomy As well,the distribution of local anesthetic within the inter-pleural space is gravity dependent.90The upright positionassumed by post-thoracotomy patients, because of itsbeneficial effects on pulmonary function, encouragespooling of the local anesthetic in the inferior thoraciccage, thereby contributing to lesser analgesia at the morecranial thoracic dermatomes Finally, the dorsal rami ofthe thoracic spinal nerves are not blocked by interpleuraltechniques; thus, patients may experience pain in themedial part of the incision and paravertebral surround-ing muscles and ligaments.64

anal-The effects of interpleural analgesia on postoperativepulmonary function are likewise unimpressive Moststudies have failed to demonstrate an improvement inFEV1, FVC, PEFR, arterial blood gas values, and/orpulmonary complications compared with these effectswhen placebo or parenteral opioids are used.74,83,84InRichardson and colleagues’ review of spirometric functionwith different analgesic techniques post-thoracotomy,

TABLE 1-6 Examples of Dosing Regimens for Interpleural Analgesia

Richardson et al, 20 mL 0.25% bupivacaine at 0.5% bupivacaine

Raffin et al, 1994 85 0.15 mL/kg 2% lidocaine with 0.05 mL/kg/h 2%

1:200,000 epinephrine after lidocaine with

infusion prn = according to circumstances.

an overall 35% preservation of function (vs the

preoper-ative values) was noted for interpleural analgesia by 48

hours postoperatively.8This was lower than for all the

other techniques examined, including intercostal nerve

blocks, thoracic paravertebral, and epidural analgesia In

two randomized studies comparing interpleural analgesia

and TPVB, analgesia for the two techniques was

equiva-lent, but patients receiving interpleural analgesia

demon-strated significantly worse FVC and FEV1values.78,91This

observation led to the suggestion that interpleural

anal-gesia may cause direct impairment of diaphragmatic and

intercostal muscle function, either by diffusion of local

anesthetic into the diaphragm and/or intercostal muscles,

with direct inhibition of their contractile function,91 or

by blockade of the phrenic nerve as it travels through the

mediastinum and/or at its terminal branches innervating

the diaphragm No studies to date have confirmed the

validity of either theory.92

advantages and disadvantages

The primary advantage of interpleural analgesia for

post-thoracotomy pain is the ease with which the technique

can be performed It is also relatively safe, and no special

monitoring is necessary for patients receiving this form

of analgesia.77

The main disadvantage of interpleural analgesia is the

lack of consistent beneficial effects on pain relief and

pulmonary function in the post-thoracotomy patient

Possible explanations for this have been discussed above

adverse effectsThe main adverse effects of interpleural analgesia for post-

thoracotomy analgesia include toxicity owing to excessive

systemic absorption of local anesthetic, blockade of the

thoracic sympathetic chain, and stellate ganglion blockade

(with an ipsilateral Horner syndrome).93 Systemic local

anesthetic toxicity is rare because plasma concentrations

usually remain below levels associated with significant

toxicity.81,94,95When administered as a bolus dose, peak

blood levels occur 5 to 30 minutes after injection

Similarly, blockade of the thoracic sympathetic chain

rarely produces clinically significant hypotension and

bradycardia This lack of hemodynamic effects has

tradi-tionally been attributed to the unilateral nature of the

sympathetic block, although Ramajoli and De Amici have

convincingly demonstrated bilateral sympathetic blockade

of the thorax and abdomen with unilateral interpleural

instillation of both 0.25 and 0.5% bupivacaine.96 Thus,

hemodynamic stability is probably due to incomplete

blockade of the upper thoracic ganglion, resulting in little

or no effect on the cardiac sympathetic fibers and allowing

compensatory vasoconstriction of the upper extremities

contraindicationsThere are no absolute contraindications specificallyrelated to the technique of interpleural analgesia Relativecontraindications include conditions in which there is ananticipated lack of efficacy, such as with pleural fibrosis,previous surgical or chemical pleurodesis, and bron-chopleural fistula or empyema; and patients for whomthe effects of high systemic blood levels of local anes-thetic may be particularly detrimental (as discussedabove under “Intercostal Nerve Block”)

Thoracic Paravertebral Nerve Block

definition and techniqueAfter its first performance in 1905 by Hugo Sellheim,TPVB enjoyed an initial period of popularity, followed by

a dramatic decline in use in the middle of the twentiethcentury.97In the past two decades, however, there hasbeen a resurgence of interest in the technique, particu-larly in Europe

TPVB is a technique whereby local anesthetic isinjected into the paravertebral space in the thoracicregion It has also been referred to as extrapleural,extrapleural paravertebral, and extrapleural intercostalanalgesia As depicted in Figure 1-2, the paravertebralspace is a wedge-shaped region adjacent to the thoracicvertebrae in the vicinity where the spinal nerves emergefrom the intervertebral foramina Its boundaries are asfollows: posteriorly, the superior costotransverse liga-ment; laterally, the posterior intercostal membrane; ante-riorly, the parietal pleura; and medially, the posterolateralaspect of the vertebrae, intervertebral disk, and interver-tebral foramen The origin of the psoas muscle forms theinferior boundary of the paravertebral space; thus, spread

of local anesthetic below T12 is uncommon The cranialboundary of the paravertebral space has not been

Subserous fascia

Sympathetic chain

Interpleural space

Extrapleural compartment

Subendothoracic compartment Intercostal nerve

Posterior primary rami Superior costotransverse ligament

Right lung

Left

Azygos vein

Esophagus

Thoracic duct Descending aorta Pleura

Visceral Parietal Endothoracic fascia

FIGURE 1-2 Anatomy of the thoracic paravertebral space.

Reproduced with permission from Karmaker MK 98

defined, and radiocontrast dye has been observed in the

cervical region after thoracic paravertebral injection.98

The thoracic paravertebral space is in continuity with the

epidural space medially via the intervertebral foramen,

the intercostal space laterally, and the contralateral

paravertebral space via the prevertebral and epidural

spaces.98The paravertebral space is traversed by the

inter-costal nerves, their dorsal rami, the rami communicantes,

and the sympathetic chain

As with other techniques, TPVB may be performed by

direct injection through a needle or an indwelling

catheter, both of which may be introduced either

percuta-neously or under direct vision before the chest is closed

Sabanathan and colleagues have described a technique for

use during thoracotomy that involves reflecting the

pari-etal pleura from the posterior wound margin onto the

vertebral bodies to form an extrapleural pocket.99A

percu-taneously placed catheter is then placed into this pocket

and positioned under direct vision so that it lies against

the angles of the exposed ribs Richardson and Lonnqvist

have employed combined techniques whereby a

percuta-neously placed catheter is inserted before the surgery

begins and a bolus dose of local anesthetic is administered

to provide intraoperative anesthesia.9 7Before chest

closure, methylene blue is injected through the catheter,

and if the spread of dye is not optimal, the catheter is

reinserted by the surgeon Video-assisted placement of a

paravertebral catheter during thoracoscopy has also been

reported.100

Table 1-7 presents various dosage regimens for TPVB

Continuous infusion of local anesthetic through a

paravertebral catheter provides better pain control than

do intermittent bolus injections.101

mechanism of actionTPVB produces analgesia by blockade not only of the

intercostal nerves but also of their dorsal rami and the

sympathetic chain Owing to the continuous nature of

the paravertebral space, local anesthetic applied at one

level spreads to multiple contiguous dermatomes Using

15 mL 0.5% bupivacaine, Cheema and colleagues

demonstrated a somatic sensory block extending for a

mean of 5 (range 1 to 9) dermatomes, and a sympatheticblock over an average of 8 (range 6 to 10) dermatomes.102

However, the extent of spread is variable, as is evidenced

by these large ranges; thus, it may be necessary toperform injections at more than one site to reliably anes-thetize more than three to four segments A smallamount of local anesthetic may also exit the interverte-bral foramina to enter the epidural space, but whetherthis contributes significantly to the analgesic effects ofTPVB is questionable.98

efficacyThe efficacy of TPVB for post-thoracotomy pain controlhas been well established Lower pain score and opioid-sparing effects have been noted in several studies compar-ing TPVB with placebo and parenteral opioids,103–106

although supplemental opioids were often still necessary

In comparison to epidural blockade with local anestheticsand/or opioids, TPVB has frequently demonstrated simi-lar or better pain relief, accompanied by less nausea,vomiting, hypotension, and urinary retention.107–110

Most studies have demonstrated a significant ment of post-thoracotomy pulmonary dysfunction withTPVB compared with placebo or parenteral opioids, asdemonstrated by higher FEV1, FVC, and/or PEFRvalues.104,105,106,111 In Richardson and colleagues’ review ofvarious techniques for post-thoracotomy analgesia, TPVBdemonstrated the best preservation of pulmonary func-tion.8FEV1, FVC, and/or PEFR values had all returned toapproximately 75% of their preoperative value by 48 hourspostoperatively in patients who had received TPVB WhenTPVB has been compared directly with thoracic epiduralanalgesia, most studies have demonstrated similar effects

improve-on pulmimprove-onary functiimprove-on for the two techniques,109,110

although TPVB was associated with higher values of PEFRand oxygen saturation as measured by pulse oximetry(SpO2) in one study by Richardson and colleagues’group.1 0 7 As noted previously (see “InterpleuralAnalgesia”), TPVB has been demonstrated to produceboth better and similar effects on pulmonary function testswhen directly compared with interpleural analgesia.70,112

Likewise, there is a limited quantity of evidence thatTABLE 1-7 Examples of Dosage Regimens for Thoracic Paravertebral Blockade

or 15 mL 0.375% bupivacaine loading dose,

then 5 mL/h 0.375% bupivacaine infusion Barron et al, 1999 105 0.3 mL/kg 1% lidocaine before chest closure or 0.1 mL/kg/h 1% lidocaine infusion or 0.1 mL/kg/h

0.3 mL/kg 0.25% bupivacaine before chest closure 0.25% bupivacaine infusion Berrisford et al, 1990 111 20 mL 0.5% bupivacaine after chest closure Approximately 0.1 mL/kg/h 0.5% bupivacaine Mathews and Govenden, 1989 108 10 mL 0.25% bupivacaine after chest closure 3–10 mL/h 0.25% bupivacaine

Richardson et al, 1999 107 20 mL 0.25% bupivacaine during chest closure 0.1 mL/kg/h 0.5% bupivacaine infusion

TPVB may decrease the risk of pulmonary complications

compared with placebo and parenteral opioids

Sabanathan, Berrisford, and colleagues, in two separate

studies (with possibly overlapping subjects), have

reported fewer pulmonary complications in patients

receiving TPVB compared with placebo.104,111

TPVB has also been shown to suppress the stress

response, as measured by serum cortisol and glucose

levels, and in this respect it functioned better than

thoracic epidural analgesia.107

advantages and disadvantages

TPVB has been described as being quick and easy to

perform.98,112,113This statement should be interpreted

cautiously, however, as it was made by the main authors

regarding TPVB in the literature today, and their

experi-ences may not be applicable to other institutions This

caution may be especially relevant for centers in North

America, where TPVB is rarely taught in the

anesthesiol-ogy and surgery training programs

Other advantages of TPVB include the lack of urinary

retention and motor blockade of the lower extremities

owing to the thoracic and unilateral location of the

block.102,114 As well, the unilateral nature of the block

results in little/no direct effects on hemodynamics,102and

the doses of local anesthetic are usually less than those

associated with systemic toxicity.70Even when higher

levels have occurred, there has been no evidence of

systemic toxicity.45,110Accordingly, no special monitoring

is necessary for patients with these blocks beyond the

usual postoperative care.112As TPVB is dependent on the

use of local anesthetics for postoperative use,

opioid-related risks are theoretically avoided However,

supple-mentation with systemic opioids is often used; thus,

opioid adverse effects may be minimized but not absent

The main disadvantage of this technique is that it is

more difficult to perform than the intercostal nerve and

interpleural blocks As well, patients with a previous

thoracotomy are usually inappropriate candidates for the

block since the paravertebral space may be obliterated by

scar tissue The technique may likewise be unsuitable for

patients undergoing a pleurectomy, although successful

use of TPVB has been reported, provided the parietal

pleura covering the vertebral bodies and a few

centime-ters distally is left intact.106

adverse effectsThe incidence of adverse effects with TPVB in the post-

thoracotomy population is 10% or lower.112,115The most

frequent adverse event is hypotension,115which has been

primarily attributed to the unmasking of relative

hypo-volemia as hypotension does not occur in well-hydrated

patients receiving TPVB for the treatment of chronic

pain syndromes.98,102,112Other complications, which occurmuch less frequently, are inadvertent puncture of theepidural or subarachnoid space owing to a faulty tech-nique,97 and unilateral Horner syndrome because of thecephalad spread of anesthetic to the cervical sympatheticstructures No fatality directly related to TPVB has beenreported in the literature.98,112

contraindications

As alluded to above, a previous ipsilateral thoracotomywould be a relative contraindication to the techniquebecause of a possible obliteration of the paravertebralspace An empyema is not directly affected by manipula-tions in the paravertebral space, but the accompanyingacidosis and hyperemia may limit the efficacy of theTPVB and increase the risk of systemic absorption oflocal anesthetic Anticoagulation is a relative contraindi-cation to the technique, but the paravertebral space is lessvascular than the epidural space; thus, the risk of venouspuncture is less than with epidural analgesia As well, theconsequences of a unilateral paravertebral space hema-toma are small compared with the potentially cata-strophic consequences of an epidural hematoma.1 1 2

Similarly, TPVB is relatively contraindicated in patientswith raised intracranial pressure because of the possibil-ity of inadvertent dural puncture and subsequent brain-stem herniation However, the risk of puncture is lessthan with epidural analgesia, so in this situation, TPVBwould be the best choice of the two techniques

Epidural Analgesia

definition and technique

Epidural analgesia refers to the technique of injecting

analgesic medication into the epidural space, ing the spinal cord As with most of the other regionaltechniques discussed heretofore, epidural analgesia isalmost exclusively administered via an indwellingcatheter when used for post-thoracotomy pain relief.Similar effects may be achieved by injecting analgesicmedication into the subarachnoid space (albeit withlower doses),116 but the technique is rarely used in theUnited States because of concerns with introducing acatheter into this space The intimate proximity of thesubarachnoid space to the spinal cord poses a risk ofinjury to the spinal cord, and an association between thedevelopment of cauda equina syndrome and subarach-noid microcatheters (also known as spinal micro-catheters) has been suggested.117

surround-Local anesthetics and opioids are the two main classes

of drugs used for epidural analgesia in post-thoracotomypatients Other drugs that have been used in the epiduralspace, either alone or as adjuncts, are discussed later (see

“Other Agents”) In the early 1980s, epidural morphine

was popular, primarily because of its hemodynamic

stability compared with epidural local anesthetics and its

relatively long duration of action.5The latter permitted

bolus dosing on an as-needed basis every 6 to 24 hours

The risk of respiratory depression and slow onset of

action with epidural morphine promoted the search for

alternative opioids,118thus leading to the use of more

lipophilic epidural opioids such as fentanyl and its

analog, sufentanil Owing to the short duration of action

of these opioids, continuous infusions are necessary

Most recently the synergistic effects of combining

local anesthetics and opioids in the epidural space have

been recognized.119This synergism has been attributed to

the facilitation of opioid transport from the epidural

space to the subarachnoid space by local anesthetic,120

and production of a conformational change in the spinal

 opioid receptor by local anesthetic agents, such that

opioid binding is facilitated.121 Accordingly, continuous

infusions of opioid–local anesthetic combinations have

become popular, with the goal of providing similar or

improved analgesia with lower doses of both agents, so

that the incidence of adverse effects is reduced Although

similar or improved analgesia has been achieved in

several studies,122–126 a reduction in adverse effects has not

been universally accomplished (see “Effects Related to

Injection of Epidural Local Anesthetic–Opioid

Combinations,” below) Current literature suggests that

the combination of 10 to 12.5 µg/mL fentanyl (or

1 µg/mL sufentanil) and 0.1 to 0.125% bupivacaine is

closest to the ideal for post-thoracotomy patients,

producing a maximum of pain relief and minimum of

side effects.6,127Of interest, the addition of bupivacaine

does not seem to improve analgesia when added to

epidural meperidine.128This may be because meperidine

has significant local anesthetic properties itself and has

even been used as the sole anesthetic for lower

abdomi-nal surgery when administered in the subarachnoid

space.129Table 1-8 presents several examples of epidural

analgesia regimens

There is controversy as to whether epidural catheters

should be inserted into the thoracic or lumbar region for

thoracotomy patients Owing to the proximity of the

spinal cord to the epidural space in the thoracic region

and the greater technical difficulty of entering the

epidural space at this level of the spinal column, many

anesthesiologists are hesitant to insert a thoracic epidural

catheter They are supported by evidence that equivalent

analgesia may be achieved by lumbar and thoracic

epidural injections in post-thoracotomy patients.120,130–134

In contrast, advocates of thoracic epidural catheters

emphasize that higher volumes and/or higher doses of

epidural opioids and/or local anesthetics were needed

with the lumbar route to produce equivalent analgesia inmany of these studies, thereby suggesting that the lumbarroute may be acceptable but not optimal As well, there is

no evidence that complication rates are higher withthoracic than with lumbar epidural catheters,135 andmany of the potential advantages of epidural analgesiadiscussed below rely on blockade of the cardiac sympa-thetic fibers at T1 to T5, which is more easily accom-plished with a thoracic than with a lumbar epiduralcatheter With these considerations in mind, theapproach at our institution is to preferentially place athoracic epidural catheter; however, a high lumbarcatheter is used if this is unsuccessful

mechanism of actionThe mechanism of action of epidural opioids and localanesthetics differs Local anesthetics applied to theepidural space act primarily by blockade of nerveimpulse conduction in the axonal membrane of thespinal nerve roots as they traverse the epidural space.136

Diffusion of local anesthetic into the long tracts of thespinal cord may further contribute to the analgesiaproduced by epidural local anesthetics The various types

of nerve fibers exhibit differential sensitivity to localanesthetics: sympathetic fibers are the most easilyblocked, and motor fibers are the most resistant.1 3 7

Consequently, the concentration of local anesthetic is theprimary determinant of the depth of blockade, withhigher concentrations producing more motor blockade.The actual extent of blockade along the spinal canal

TABLE 1-8 Examples of Dosing Regimens for Epidural Analgesia

prn Sufentanil 1 µg/mL 0.1–0.2 µg/kg/h 5–7 µg q10–15 min

Data from University of Texas M D Anderson Cancer Center protocol and DeLeon-Casasola OA and Lema M 154

prn = according to circumstances.

depends primarily on the volume administered, and

because of the greater sensitivity of the sympathetic

fibers, the extent of sympathetic blockade may be greater

than the somatic sensory block.

Epidural opioids exert their primary therapeutic

effects by binding to specific opioid receptors in the

substantia gelatinosa of the dorsal horn of the spinal cord

gray mater.1 3 6 , 1 3 8 This region contains interneurons

involved in the ascending pain pathways (the

spinothala-mic and spinoreticular tracts) Opioid receptors are

located both presynaptically and postsynaptically, and

they function to inhibit the release of neurotransmitters

from primary sensory neurons and block the

depolariza-tion of post-synaptic neurons, respectively.44 The term

selective spinal analgesia has been used to denote

analge-sia attributable to these spinal cord opioid receptors

Before reaching the spinal cord, opioids injected into the

epidural space must first travel through the dura mater,

subdural space, arachnoid mater, subarachnoid space

(containing the cerebrospinal fluid), and pia mater The

epidural space contains an abundance of fat tissue and an

extensive venous plexus, and 90 to 97% of an injected

dose is absorbed into these compartments,139–141thereby

never reaching the subarachnoid space

Epidural opioids may also produce analgesia at a

supraspinal level (termed supraspinal analgesia) by

bind-ing to opioid receptors in the brain Opioids gain access

to these sites via two main pathways: absorption into the

epidural veins and subsequent entry into the systemic

circulation, and rostral travel through the cerebrospinal

fluid to the brain After a bolus injection of all epidural

opioids, plasma levels peak at approximately the same

time as with an intramuscular injection,139,142,143and in

some studies have achieved values high enough to

contribute to analgesia.139,142,144–146Plasma opioid levels fall

quickly, however, and are of little importance beyond the

first hour after epidural bolus administration for all

agents.139,142,146

A different scenario arises when lipophilic agents

(such as fentanyl and sufentanil) are administered by

continuous epidural infusion or repeat bolus

Continuing systemic absorption of these agents results

in accumulation, and some studies have recorded

systemic plasma levels within the usual therapeutic

range for these drugs, when administered by these

meth-ods.123,143 Similarly, Miguel and colleagues and Sandler

and colleagues have demonstrated that epidural

infu-sions of fentanyl and sufentanil produce plasma levels

similar to those with intravenous infusion, when titrated

to equivalent analgesia.1 4 7 , 1 4 8 This suggests that

supraspinal analgesia may be a major contributor to the

overall analgesic effect when lipophilic opioids are

administered in this manner

Rostral travel through the cerebrospinal fluid ofopioids injected into the epidural space is most promi-nent with morphine, as its relative hydrophilic propertieslimit its diffusion out of the subarachnoid space, therebyallowing greater quantities of morphine to be retained inthe cerebrospinal fluid for a prolonged period of time.149

Morphine travels cranially via the slow process of brospinal fluid bulk flow, leading to peak levels ofmorphine in the cervical cord region by 3 to 5 hours afterlumbar epidural injection.136,150

cere-Movement through thecerebrospinal fluid for more lipophilic opioids may alsooccur, especially with large bolus doses, but the quantities

of drug detected in the cervical cord and/or cisternamagna have been small, and their contribution to theanalgesia achieved with these agents is unknown.146,151

efficacyThe efficacy of epidural analgesia in providing painrelief after thoracotomy depends on the drug(s) used.Epidural analgesia with local anesthetics alone is moreeffective in providing analgesia than are parenteralopioids, but the concentrations needed to accomplishthis (eg, 0.5% bupivacaine) are accompanied by a signif-icant risk of hypotension When lower concentrationshave been used, supplemental parenteral opioids areusually necessary.70,107,152

The efficacy of epidural morphine in providingpost-thoracotomy analgesia is undisputed It is consid-ered the “gold standard” for epidural opioid analgesia.Pain scores and/or the need for supplemental anal-gesics are universally lower for epidural morphine thanfor parenteral morphine, and these effects are accom-plished using lower doses of epidural morphine, whichlast longer than parenteral morphine.5 , 1 5 2 , 1 5 3 Thelipophilic opioids are also effective in providing analge-sia after thoracotomy when administered via theepidural route However, as discussed previously, there

is evidence that continuous epidural infusions of theseagents produce post-thoracotomy pain relief primarily

by the systemic absorption of the opioid and may offerlittle advantage over the less complicated intravenousroute of administration.154

The opioid–local anesthetic combinations populartoday are also very effective in providing pain relief afterthoracotomy As combinations are relatively new tech-niques and the efficacy of epidural analgesia for post-thoracotomy pain has already been established, there hasbeen little interest in performing studies comparing theefficacy of combinations to that of parenteral opioids orplacebos Nevertheless, improved analgesia has beennoted with both epidural morphine–bupivacaine andepidural fentanyl–bupivacaine infusions compared withparental opioids.152,155

Studies comparing epidural analgesia with other

modes of regional analgesia in post-thoracotomy patients

are few, and their interpretation has been confounded by

the use of a variety of different medications in the

epidural space Three studies have used epidural local

anesthetics alone Brockmeier and colleagues showed no

difference in analgesic efficacy between 0.375% epidural

bupivacaine and interpleural analgesia.86Richardson and

colleagues demonstrated better analgesia with TPVB

than with epidural analgesia, but the TPVB group

received 0.5% bupivacaine and the epidural group

received only 0.25% bupivacaine.1 0 7 A final study

compared 0.25% epidural bupivacaine with 0.25% TPVB

bupivacaine and 0.5% interpleural bupivacaine.70All

techniques produced similar analgesia at rest, but the

intercostal nerve blocks group had better dynamic pain

relief for the first 4 hours after thoracotomy In a study of

epidural analgesia using a combination of fentanyl and

bupivacaine, analgesia was superior to that produced by

TPVB,114although this effect did not persist beyond the

first postoperative day

Evidence regarding the efficacy of epidural analgesia in

improving pulmonary function and decreasing pulmonary

morbidity in the post-thoracotomy patient is conflicting

Many studies have revealed no difference in arterial blood

gas results, spirometry measurements, or pulmonary

complications when epidural analgesia has been compared

with parenteral opioids or other types of regional analgesia

in this population.6 9 , 8 8 , 1 0 9 , 1 5 3 , 1 5 6 – 1 6 0 In Richardson and

colleagues’ review of different techniques for

post-thoraco-tomy analgesia discussed previously, epidural analgesia

with local anesthetics and/or opioids resulted in a

moder-ate preservation of pulmonary function.8By 48 hours after

surgery, FEV1, FVC and/or PEFR values had returned to

approximately 55% of their preoperative values in patients

who had received epidural analgesia, which was similar to

the results obtained with intercostal nerve blocks but

worse than the 75% values observed with TPVB

For those studies that have shown an improvement in

pulmonar y parameters, most have demonstrated

improvement in only some of the parameters measured

For example, Guinard and colleagues demonstrated

higher FVC and FEV1 values for thoracic epidural

fentanyl when compared with intravenous fentanyl, but

no difference in arterial blood gas results or the number

of patients with abnormalities on chest radiographs.132

Salomaki and colleagues have shown lower PaCO2values

with epidural fentanyl but similar PaO2and incidences of

atelectasis compared with intravenous fentanyl.161Two

articles from Hasenbos and colleagues are the only

stud-ies that have found both improved arterial blood gases

(PaCO2 less elevated above preoperative levels) and

reduced incidence of pulmonary complications.162,163

However, these studies were not blinded, and the opioidexamined was nicomorphine, the 3,6-dinicotinoyl ester

of morphine, which is not available in North America Aswell, the sole analgesic in the nonepidural groups wasintramuscular nicomorphine, administered in as-neededdoses by the nursing staff By providing parenteral opioid

in this manner, analgesic therapy in these groups may nothave been optimized

A recent meta-analysis of the pulmonary effects ofvarious analgesic regimens in a wide variety of surgicalprocedures (including but not restricted to thoraco-tomies) revealed only a diminished incidence of atelecta-sis with epidural opioids and a decreased incidence ofpulmonary infection and overall pulmonary complica-tions, plus an increased PaO2with epidural local anes-thetics.164Of interest, the authors emphasize the lack ofdifference in spirometry results between the differentmethods of analgesia and suggest that there is no ratio-nale for using these surrogate measures of pulmonaryoutcomes

Epidural analgesia has little, if any, impact on ing the stress response to surgery in the post-thoracotomypopulation.107,165This has been attributed to incompleteblockade of the afferent sensory nervous input from thesite of surgery and the release of components of the stressresponse, such as cytokines, directly into the bloodstreamfrom the site of tissue injury.166

modify-advantages and dismodify-advantagesOne major advantage of epidural analgesia is related tothe use of opioids in the epidural space Systemic absorp-tion and/or cephalad spread of epidural opioid may alle-viate the shoulder pain commonly associated withthoracotomies, and even neck incisions for esophagec-tomies In our institution, we do not routinely use supple-mental parenteral opioids for pain in these two locations

If necessary, NSAIDs and acetaminophen are almostalways sufficient adjuvants to our epidural analgesia.Thoracic epidural analgesia with local anesthetics (alone

or in combination with opioids) may have unique tages in patients with coronary artery disease Blockade ofthe cardiac sympathetic fibers innervating the heart(T1–T5) results in small reductions in heart rate, systemicvascular resistance, and possibly cardiac output,167,168therebydecreasing myocardial oxygen demand At the same time,myocardial oxygen supply may improve, particularly inareas at most risk of ischemia Thoracic epidural analgesiawith local anesthetic has been demonstrated to producedilatation of stenotic coronary arteries,169redistribution ofblood flow from the epicardium to endocardium,170andredistribution of blood flow specifically toward ischemicregions of the myocardium.170Maintaining the systemicblood pressure close to the normal range (eg, mean arterial

advan-pressure < 20% below baseline) is necessary for these

effects to be most evident.171

Another potential benefit of epidural analgesia in the

patient with coronary artery disease relates to coagulation

Local anesthetic may be absorbed from the epidural space

in quantities sufficient to interfere with platelet

aggrega-tion,172,173thereby counteracting the hypercoagulable state

associated with major surgery and potentially diminishing

the risk of coronary artery thrombosis formation

Despite the above observations, there have been no

properly conducted randomized controlled trials

demon-strating decreased risk of myocardial ischemia/infarction

through the use of epidural analgesia in any group of

patients postoperatively, let alone those having

thoraco-tomies As well, there is concern that blockade of the

sensory innervation of the upper thoracic region may

simply obliterate the pain of myocardial ischemia, thus

removing an important warning signal of impending

myocardial infarction.174

Epidural analgesia may also decrease the risk of

postop-erative arrhythmias This is of particular relevance in the

post-thoracotomy situation as supraventricular

arrhyth-mias (especially atrial fibrillation) occur in approximately

15% of patients after lung surgery, and recurrent episodes

have been associated with increased perioperative

mortal-ity.175Animal studies suggest that thoracic epidural

analge-sia with local anesthetic may reduce the risk of ventricular

tachyarrhythmias and reentry supraventricular

arrhyth-mias,176–178an effect that has been attributed to cardiac

sympathetic blockade In humans a retrospective review by

Groban and colleagues noted a significantly lower

inci-dence of atrial arrhythmias while thoracic epidural

analge-sia (with opioid or opioid plus local anesthetic) was in use

compared with the incidence after the epidural was

removed on the second or third postoperative day.179In a

prospective, randomized study comparing thoracic

epidural bupivacaine with thoracic epidural morphine,

Oka and colleagues demonstrated a lower incidence of

supraventricular tachyarrhythmias with epidural

bupiva-caine when administered for 3 days after thoracotomy.180

The aforementioned cardiovascular benefits of

thoracic epidural analgesia with local anesthetics cannot

be extrapolated to the use of epidural opioids or to

epidural local anesthetics administered by the lumbar

route In the former instance, there is no direct inhibition

of sympathetic input to the heart In the latter, blockade

of T1 to T5 would require such an extensive blockade of

the sympathetic nervous system that the resultant

decrease in blood pressure would very likely counteract

any beneficial effects attributable to the blockade of the

cardiac sympathetic fibers

The main disadvantage of epidural analgesia is that,

of the regional techniques described heretofore, it is

probably the most difficult to perform and is almostexclusively performed by an anesthesiologist or nurseanesthetist Although many of these individuals havehad extensive training and/or experience in performingepidurals for obstetric indications, and thus can facilelyinsert an epidural in the lumbar region, thoracic epidu-rals for analgesia after thoracotomy are different Notonly is the thoracic epidural space more difficult toidentify owing to anatomic differences in the spinalarchitecture of the thoracic region, but the majority ofthoracotomy patients are elderly, with calcified supra-spinous and interspinous ligaments and compressedintervertebral spaces, all of which add to the difficulty ofsuccessfully inserting an epidural

Although rare, subdural hematoma and pneumocephalushave been reported after spinal puncture.1 8 2 Morecommonly, a medication intended for the epidural space

is injected into the subarachnoid space instead, whichmay have disastrous consequences An epidural dose ofeither local anesthetic or opioid would be clearly exces-sive in the subarachnoid space, potentially producingmajor motor blockade, hypotension and “total spinalanesthesia” with the former class of drugs, and life-threatening respiratory depression with opioids

An inadvertent spinal puncture may also produce aheadache, which can be incapacitating This postduralpuncture headache may be frontal and/or occipital,usually develops 24 to 72 hours after the spinal puncture,and is due to leakage of cerebrospinal fluid through thehole in the subarachnoid membrane and subsequenttraction on pain-sensitive structures at the base of thebrainstem.183It may be associated with nausea/vomitingand cranial nerve palsies and auditory disturbances,184,185

and it is clearly differentiated from other causes ofheadache by its prominent exacerbation with the uprightposition and complete resolution with the supine posi-tion Therapy is important, not only from a humanitar-ian point of view, but because the headache maydiscourage the patient from coughing and assuming theupright position, thereby potentially interfering with

recovery In addition to nonopioid and opioid analgesics,

more specific therapy with oral or intravenous caffeine

and aggressive fluid intake instillation is usually

success-ful in alleviating the postdural puncture headache The

latter is often undesirable in the post-thoracotomy

patient, however When these measures fail, instillation of

10 to 30 mL of normal saline or 20 mL of freshly

obtained autologous blood (an “epidural blood patch”)

into the epidural space, is indicated.186

The potentially catastrophic complications of epidural

needle or catheter insertion, epidural hematoma and

abscess, and permanent neurologic deficits are,

fortu-nately, very rare Incidence rates have been estimated to be

≤ 1 in 150,000 to 190,000,187,188≤ 1 in 1,000 to 6,500,189–191

and ≤ 1 in 3,000 to 4,500,135,192 respectively Permanent

neurologic deficits (usually paraplegia) may occur as a

consequence of cord compression and ischemia induced

by an epidural hematoma or abscess Additionally, they

may be related to mechanical trauma from the epidural

needle or catheter, inadvertent injection of a neurotoxic

substance into the epidural space, or spinal cord ischemia

from other causes, such as epinephrine-induced

vasocon-striction of arteries supplying the cord, pressure effect

from the volume of epidural injectate, and hypotension

induced by the sympathetic block.193

Epidural abscesses may arise from direct extension of

infection in the local area of the epidural insertion site,

or from infection at a remote part of the body, with

bacteremia and subsequent seeding of the epidural

space.194,195Factors associated with the occurrence of

epidural abscesses are duration of epidural

catheteriza-tion ≥ 3 days190; immunocompromise associated with one

or more complicating disease, such as cancer, acquired

immunodeficiency syndrome, diabetes, multiple trauma,

chronic renal failure, or chronic obstructive pulmonary

disease190,195,196; multiple attempts at inserting the epidural

needle/catheter191; and the use of low-dose unfractionated

or low molecular weight heparin.190The association with

the latter two factors may reflect the development of a

hematoma within the epidural space, which subsequently

acts as a nidus for infection

Direct puncture of the extensive epidural venous plexus

during epidural needle or catheter insertion may result in

the development of an epidural hematoma Although it

may initially be small and clinically unimportant, this

hematoma may later enlarge when the clot is disrupted by

a coagulation abnormality or direct trauma The latter

may occur at any time while the catheter is being used as

epidural catheters are well-known to migrate within the

spinal canal.1 8 2 However, a clot is most likely to be

disrupted when an epidural catheter is removed from the

patient Approximately half of the epidural hematomas

attributed to epidural analgesia in Vandermeulen and

colleagues’ 1994 review of the literature occurred ately upon removal of the epidural catheter.197Risk factorsassociated with the development of an epidural hematomainclude increased age,198multiple attempts at inserting theepidural needle or catheter,198and coagulation defects.188,198

immedi-It is controversial whether the appearance of blood in theneedle or catheter during epidural insertion (known as a

“traumatic” needle/catheter insertion) is a risk factor forthe development of a significant epidural hematoma.Given the extensive vascularity of the epidural space andthe 3 to 12% incidence of puncture of a blood vessel orvessels during epidural catheter insertion,182there are obvi-ously a large number of epidural hematomas that are toosmall to attain clinical importance

The association between epidural hematomas andcoagulation defects has been recognized for decades.197

Defects in either the platelet-mediated or coagulationfactor–dependent phases of the coagulation systemenhance the possibility of epidural hematoma formation,and combinations of defects increase the risk further.198–200

Therapeutic anticoagulation with heparin, warfarin, andantifibrinolytics are known offenders, as are therapeuticand prophylactic doses of low molecular weightheparin.187,197,201The use of NSAIDs and low-dose prophy-lactic unfractionated heparin has not been associatedwith a significantly increased risk of epidural hematomaformation.187,188,202

Effects Related to Epidural Injection of Local Anesthetics Adverse effects of local anesthetics injected

into the epidural space include hypotension, motor ade, urinary retention, and systemic toxicity These effectsare usually dose related, being most pronounced whenhigher concentrations of local anesthetic are used.Hypotension is a frequent occurrence as the sympatheticblockade produces peripheral vasodilation and bradycar-dia, and it is generally the limiting factor in the use of highconcentrations of local anesthetic For example, El-Bazand colleagues reported a 23% incidence of systolic bloodpressure < 60 mm Hg with heart rate < 60 beats perminute in the first 24 hours after surgery, using 5 mL bolusdoses of 0.5% bupivacaine administered on an as-neededbasis through a thoracic epidural catheter.4Motor block-ade of the legs interfering with ambulation is uncommon

block-if the epidural catheter is inserted in the thoracic region.203

In contrast, weakness of the upper extremities maydevelop with an epidural at this level and may be distress-ing to patients.4Urinary retention should not occur withthoracic epidurals if the volume of local anesthetic islimited to that necessary for blockade of just the thoracicdermatomes Toxicity owing to systemic absorption is arare occurrence at the usual recommended doses

Effects Related to Epidural Injection of Opioids.

Epidural opioid adverse effects are similar to those

antici-pated with parenteral administration of this class of

drugs The most frequent are nausea/vomiting, pruritus,

urinary retention, and intestinal hypomotility and

somnolence, and the most important is respiratory

depression Many of these effects are dose related and

occur more frequently in patients who are opioid naive.204

They are usually mediated by opioid receptors, such that

opioid antagonists should be effective in the prevention

and treatment of these effects.205Unfortunately, the doses

of antagonist needed for this purpose may overlap with

the doses that will antagonize the analgesic effects;

there-fore, other treatment modalities are usually employed

first, and opioid antagonists are reserved for instances

when they fail Compared with parenteral opioid

admin-istration, the epidural route is associated with a greater

incidence of pruritus and urinary retention, a lower

inci-dence of somnolence, and an equivalent inciinci-dence of

nausea/vomiting and respiratory depression.204 Pruritus,

nausea/vomiting, and urinary retention occur less

frequently when epidural morphine is administered as a

continuous infusion than as intermittent bolus

injec-tions.4

Respiratory Depression Respiratory depression is the

most feared complication of epidural opioids Large series

have reported 0.1 to 3% incidences of “clinically

signifi-cant” respiratory depression,118,189,203,206–208defined as that

requiring intervention These represent combined data

from patients having a wide variety of surgical procedures,

who have received different types of medication,

adminis-tered by varying protocols, through catheters inserted at

both lumbar and thoracic levels of the spinal column The

incidence appears to be similar for most opioids,209with

the possible exception of an increased incidence with high

bolus doses of sufentanil.210After a single bolus dose of

epidural opioid, respiratory depression classically follows

two distinct patterns, which have been designated early

and late.204Early respiratory depression occurs within the

first 2 to 4 hours after administration of the epidural

opioid, has been observed with most opioids used during

epidural analgesia, correlates with high peak plasma levels,

and is thus thought to be primarily due to systemic

absorption of the opioid Late depression occurs at > 2 to 4

hours after a dose of epidural opioid—most often at 6 to

12 hours204,211—but there have been reports of respiratory

depression persisting as long as 22 hours after

administra-tion of a bolus dose of epidural opioid.205,212,213This type of

respiratory depression has been evidenced almost

exclu-sively with morphine and is due to the rostral spread of

opioid through the cerebrospinal fluid to the respiratory

control centers in the brainstem With the frequent use of

infusions in recent years, the concepts of early and late

have become obscured, and respiratory depression may

occur at any time

Respiratory depression with epidural opioids usuallypresents with a slow respiratory rate, although there havebeen reports of severe hypercarbia with a normal respira-tory rate.214,215In these examples, somnolence was present;hence, there is a need for the assessment of the level ofconsciousness as an indicator of respiratory depression.Well-established risk factors for the development ofrespiratory depression include older age, AmericanSociety of Anesthesiologists physical status classes III to

IV, respiratory disease, sleep apnea syndrome, elevatedintrathoracic pressure, and concomitant use of systemicopioids and/or other central nervous system depres-sants.118More controversial risk factors are bolus injec-tions, compared with continuous epidural infusions154,216;and thoracic epidural catheters, in contrast to lumbarepidurals.118,130The latter has been demonstrated onlywith morphine

Efforts to prevent postoperative respiratory depressionassociated with epidural opioids have focused on usingthe minimally effective dose of epidural opioid, limitingthe quantity of all opioids and sedatives used intraopera-tively, and avoiding concomitant use of parenteral opioidsand other central nervous system depressants.214Frequentmonitoring of patients for evidence of respiratory depres-sion is necessary, most often with intermittent assessment

of respiratory rate and level of consciousness every 0.5 to

2 hours, plus continuous pulse oximetry Reliance onpulse oximetry alone is not recommended as a decrease inoxyhemoglobin saturation may be a late sign of respira-tory depression when supplemental oxygen is used.Arterial blood gases and/or continuous ventilatory moni-tors may be considered for high-risk patients.2 1 5

Monitoring should be applied throughout the course ofadministration of the epidural opioids and continued for

4 to 6 hours after the last dose or after stopping an sion.118,204Morphine is the exception as the risk of delayedrespiratory depression mandates a more protractedperiod of monitoring: 12 to 24 hours after the last dosehas been recommended.118,182,204,205,214The location of themonitoring has changed over the last few years In the1980s it was recommended that all patients with epiduralopioids (with the possible exception of the obstetricpopulation) be observed in a setting such as an intensivecare unit or postanesthesia care unit.118 More recently,large series have demonstrated the apparent safety ofcaring for these patients on a ward, provided the staff iswell educated and the aforementioned level of monitoring

evidence for this varies between studies Most

consis-tently, the incidence and severity of hypotension is less

than that associated with local anesthetic alone,126,171,218

and the incidence of somnolence is lower than that

asso-ciated with epidural opioids alone.123At our institution

we usually initiate epidural analgesia therapy with a

solu-tion containing fentanyl 10 µg/mL and bupivacaine

0.075% and subsequently adjust the two drugs

indepen-dently, in accordance with the adverse effect profile of the

individual patient

contraindicationsRelative contraindications to the use of epidural analge-

sia include elevated intracranial pressure, preexisting

disease of the spinal cord or peripheral nervous system,

infection, and coagulopathy Elevated intracranial

pres-sure is listed here because of the risk of inadvertent

dural puncture and subsequent brainstem herniation

With neurologic disease, there is concern that the

epidural analgesia may exacerbate the underlying

disease, as has been suggested for demyelinating diseases

such as multiple sclerosis,219or that a coincidental

deteri-oration in neurologic function may be incorrectly

attrib-uted to the epidural

Owing to the potential risk of epidural abscess

forma-tion, infection at the intended site of epidural insertion is

considered an absolute contraindication, and evidence of

infection at a more distant site is a relative

contraindica-tion to the use of epidural analgesia Although no studies

have properly addressed this issue, epidural analgesia is

probably acceptable if the distant infection is completely

localized, but it is generally recommended that epidural

analgesia be avoided in the possible presence of

bacteremia, as evidenced by an elevated white blood cell

count or temperature The situation becomes more

controversial if appropriate antibiotics have been started

The literature also does not address the issue of epidural

analgesia in the patient with an empyema but no evidence

of systemic infection, but many anesthesiologists would

be reluctant to insert an epidural in this scenario

Further-more, the question of whether the catheter should be

removed when an infection develops in a patient with anindwelling epidural catheter also has not been discussed

in the literature At our institution we approach each case

on an individual basis, taking into consideration suchfactors as the duration of epidural use, whether thepatient appears to be benefiting from the epidural, andwhether appropriate antibiotics have been initiated

To diminish the risk of epidural hematoma formation,epidural analgesia is usually avoided in patients with apreexisting coagulopathy It has traditionally been taughtthat the prothrombin time, activated partial thrombo-plastin time (aPTT), and international normalized ratio(INR) should all be within the normal range and thatplatelets should number ≥ 100,000/mm3 before anepidural needle or catheter is inserted.220More recentrecommendations have suggested that epidural analgesiamay be safely considered with an INR up to 1.4 andplatelet counts as low as 80,000/mm3.197 Bleeding timesare no longer advocated as a reliable method of deter-mining platelet function.187If patients have been on anymedications effecting coagulation preoperatively, thesedrugs should ideally be stopped before the epidural isinserted, and the epidural catheter should be removedbefore they are restarted Guidelines for specific medica-tions are presented in Table 1-9

A more difficult situation arises when medicationsinterfering with coagulation are administered in the peri-operative period The most conservative approach is tocompletely avoid epidural analgesia in these patients Ininstitutions with the expertise to provide good post-thoracotomy analgesia by other methods (especially TPVBand continuous intercostal nerve blockade), this is areasonable approach Postoperative use of NSAIDs andsubcutaneous unfractionated heparin appears to be safewhen used in conjunction with epidural analgesia, based

on the very small number of case reports of patients oping epidural hematomas in these circumstances and thelack of hematoma development in the few larger studiesthat have addressed this issue.198,221However, the response

devel-to subcutaneous unfractionated heparin can be able,222,223and patients with cachexia and/or liver dysfunc-

vari-TABLE 1-9 Epidural Management in Patients Receiving Agents Affecting Coagulation Preoperatively

Subcutaneous unfractionated heparin (1,000 U q8–12h) Insert EP ≥ 4 h after last dose aPTT if cachectic or liver dysfunction; platelets if heparin

taken for > 4 d

Thrombolytics and fibrinolytics (eg, streptokinase, t-PA) Insert EP ≥ 1–2 d after last dose Fibrinogen

Data from Horlocker TT, 187 Vandermeulen EP et al, 197 Horlocker TT et al, 198 Tryba M, 201 Heit JA 254

aPTT = activated partial thromboplastin time; EP = epidural; INR = international normalized ratio; LMWH = low molecular weight heparin; PT = prothrombin time; t-PA = tissue plasminogen activator.

tion may be particularly sensitive to the anticoagulant

effects of the drug As well, use of any form of heparin

beyond approximately 4 days may lead to

thrombocytope-nia.201It is therefore most prudent to monitor the aPTT

and platelet count regularly in these patients Guidelines

for the use of other agents are summarized in Table 1-10

If one decides to perform epidural analgesia in

patients at risk of developing an epidural hematoma or

abscess, frequent assessment of neurologic function

(optimally every 2 hours187) is essential As well, the

concentrations of local anesthetics in the epidural

solu-tion should be minimized, or local anesthetics avoided

entirely, so that if a neurologic abnormality is detected,

there is no confusion between drug effect and actual

pathology Immediate diagnosis of an epidural

hematoma is essential as permanent neurologic deficit is

more likely if surgical decompression is delayed beyond

the first 8 to 12 hours after presentation.197Although it is

controversial whether blood in the epidural needle or

catheter is a risk factor for the development of

hema-toma, the policy at many institutions is to cancel or delay

surgery for up to 24 hours if blood is noted during

needle or catheter insertion, and systemic heparinization

is planned.187,197Another approach, employed by Loik and

colleagues, for patients undergoing cardiac surgery has

been to routinely insert the epidural catheter the day

before surgery, thereby maximizing the time interval

between catheter insertion and heparinization.224

durationThe optimal duration for epidural analgesia after thoraco-

tomy has not been established Depending on the

institu-tion, epidural catheters may be removed at 48 to 72 hours

postoperatively or continued for up to 7 days, or even

longer As much of the pain after thoracotomy is

attribut-able to the chest tubes, the general guideline at our

insti-tution is to maintain the epidural until these tubes are

removed Thus, epidural catheters usually remain in place

for 3 to 7 days postoperatively Pneumonectomy patients

are an exception to the “chest tube rule” as these tubes aregenerally removed 24 hours postoperatively but epiduralanalgesia is continued for at least 3 days At the otherextreme, patients with esophagectomies often have theirchest tubes continued for more than a week As epiduralmedication requirements are usually minimal by 7 daysafter this surgery, we generally remove the catheter on theseventh operative day in these patients

The only study that has addressed this issue is byNomori and colleagues, who performed a retrospectivereview of data from patients having an anterior limitedthoracotomy who received epidural morphine bycontinuous infusion for either 3 or 8 days postopera-tively.23 There were no differences between groups withrespect to pulmonary function tests or pain scores forthe first 7 postoperative days, but the day 8 group hadmore pain than did the day 3 group on the eighth andninth postoperative days As well, the pain scores on day

8 after the epidural was removed in the day 8 group weresignificantly higher than the pain scores on the seventhpostoperative day By contrast, the day 3 group experi-enced no such “rebound” after their catheters wereremoved These observations suggest that, at least for ananterior limited thoracotomy, extending the use ofepidural analgesia beyond the first 3 postoperative dayshas no benefit and may even exert a negative impact onoverall pain control Despite the limitations of the studydesign, the results of this study are intriguing, and futurerandomized controlled trials are warranted to establishthe optimal duration of epidural analgesia from a risk-benefit perspective

other agents

In addition to local anesthetics and opioids, many otheragents have been examined as potential candidates forepidural analgesia These include clonidine, opioid agonist-antagonists, ketamine, verapamil, and methylprednisolone.Epidural clonidine has been used for postoperativeanalgesia in a variety of procedures Its mechanism ofTABLE 1-10 Epidural Management in Patients to Receive Agents Affecting Coagulation Perioperatively*

Epidural Catheter Subcutaneous unfractionated Start ≥ 1 h after EP insertion Remove ≥ 4 h after last dose aPTT if cachectic or liver dysfunction;

and wait ≥ 2 h after removal before

injecting next dose Intravenous heparin Start ≥ 1 h after EP insertion Remove ≥ 2–4 h after last dose aPTT; platelets if heparin taken for > 4 d

Data from Horlocker TT, 187 Vandermeulen EP et al, 197 Horlocker TT et al, 198 Tryba M, 201 and Enneking FK, and Nenzon H 255

aPTT = activated partial thromboplastin time; EP = epidural; INR = international normalized ratio; LMWH = low molecular weight heparin; PT = prothrombin time.

*The use of EP catheters in patients who will receive thrombolytics and fibrinolytics is strongly discouraged.

action is primarily by stimulation of2-adrenergic

recep-tors in the dorsal horn of the spinal cord,58although

clonidine also produces direct blockade of conduction in

A delta and C nerve fibers The main advantage of

epidural clonidine relates to its lack of most of the classic

opioid-related adverse effects, such as nausea/vomiting,

pruritus, and respiratory depression However, epidural

clonidine may produce sedation, hypotension, and

bradycardia owing to stimulation of intracranial 2

-adrenergic receptors secondary to systemic absorption of

the drug In nonthoracic surgery, clonidine is an effective

analgesic when administered alone in the epidural space,

but the doses required are high enough to produce a

significant incidence and severity of these

clonidine-related adverse effects.225,226Combinations of clonidine

and local anesthetics or opioids produce additive, and

possibly synergistic, analgesic effects, resulting in

improved analgesia with relatively low doses of

cloni-dine.227,228However, sedation and mild hypotension have

still been a problem in these studies In the single

investi-gation focusing on post-thoracotomy patients, a single

bolus of epidural clonidine exhibited no effect on pain

relief compared with placebo.229 Further studies are

necessary in this population

Medications with opioid agonist-antagonist properties

have been used for epidural analgesia with the goal of

reducing the opioid-related side effects, especially

respi-ratory depression In the only two studies focusing on

post-thoracotomy patients, lumbar epidural nalbuphine

provided significantly less analgesia than did lumbar

epidural morphine.230,231It is unknown whether a thoracic

approach would have produced different results In a

group of 20 gynecologic patients, epidural pentazocine

produced good postoperative analgesia, with 16 patients

being completely pain free and the other 4 reporting mild

pain but requiring no additional analgesics.232A further

benefit in this study was the lack of urinary retention in

all 16 epidural pentazocine patients who did not have an

indwelling urinary catheter Accordingly, this drug

appears to be promising but needs to be examined in the

post-thoracotomy population

Discovery of the integral role of NMDA receptors in

processing nociceptive information in the spinal cord has

led to much interest in ketamine as an epidural analgesic

agent.233Ketamine is a potent, noncompetitive inhibitor

of the NMDA receptor and has the major clinical

advan-tage of exhibiting no respiratory depressant effects

Epidural ketamine has not been studied in the

post-thoracotomy population, but for other types of surgery

ketamine, appears to have little or no analgesic efficacy

when used as a sole agent in the epidural space.234–236More

promising results have occurred when epidural ketamine

has been combined with epidural opioids and/or local

anesthetics In many of these investigations, ketaminepotentiated the analgesic effects of the other agents.237,238

A disadvantage of ketamine relates to its propensity tocause psychomimetic effects, which have been reported

in up to 15% of patients receiving ketamine by theepidural route.239These effects have been successfullytreated with discontinuation of the drug and/or withsystemic benzodiazepines

Verapamil has been advocated as a possibly usefulepidural analgesic agent based on observations thatneurotransmitter release is mediated by calcium influxinto the synaptic terminals of neurons.240By interferingwith this influx, normal sensory processing in the spinalcord is disrupted, and verapamil has been shown topotentiate the antinociceptive effects of morphine at thespinal cord level in an animal model.239In a preliminarystudy of patients having lower abdominal surgery, vera-pamil added to an epidural bupivacaine solutiondecreased supplemental opioid requirements.241Furtherinvestigations are necessary to establish the safety ofverapamil in the epidural space from a perspective ofneurotoxicity and to examine the analgesic efficacy ofepidural verapamil in the post-thoracotomy population.Investigations of patients having spinal surgery havedemonstrated moderate effectiveness of epidural glucocor-ticoids as analgesic agents.241This has been primarilyattributed to localized anti-inflammatory activity of thesedrugs, but evidence that glucocorticoids may also influ-ence synaptic transmission in the spinal cord has led to thesuggestion that epidural glucocorticoids may be useful forother types of surgery as well.242Blanloeil and colleaguesexamined the effects of a continuous infusion of epiduralmethylprednisolone on pain relief after posterolateralthoracotomy, but they found no difference in analgesiabetween patients who received methylprednisolone andthose who received placebo.243Accordingly, there is noevidence at present that epidural glucocorticoids will beuseful agents for post-thoracotomy analgesia

Management of Pain

The management of postoperative pain in general haslong been the sole responsibility of the surgeon after theacute phase in the postanesthesia care unit Probablyconcomitantly with the introduction of epidural analge-sia, anesthesiologists have undertaken some sharedresponsibility for the control of post-thoracotomy pain.This has improved postoperative analgesia and arguablyimproved intraoperative management Whether a formal

“thoracic team,” which incorporates representatives fromthoracic surgery, anesthesiology, nursing, and painmanagement, improves outcome is debatable Improvedoutcome has only been suggested for pediatric fellowship

training,244and the mechanism of the improved outcome

still may just be due to more opportunity to maintain

skills.245 It is likely, however, that the simple existence of a

group of identifiable individuals with an interest in pain

relief both during and after thoracic surgery would

enhance immediate feedback within the group This

would predispose to rapid adoption of needed

improve-ments and facilitate adherence to the principles of

continuous quality improvement Familiarity and

repeti-tion should increase the success rate of procedures and

the satisfaction felt by patients

Certainly epidural anesthesia should be managed

solely by an anesthesia provider due to the specialized

knowledge required for implementing and safely

provid-ing this service Most large hospitals have an acute pain

service or postoperative pain service in addition to the

anesthesiologists in the operating rooms The

postopera-tive pain service not only helps manage the ongoing

needs of patients with epidural analgesia, but is also

available to manage or consult on other postoperative

analgesia–related dilemmas In addition, almost all

medical communities have available chronic pain

special-ists or even a chronic pain service Specialspecial-ists in chronic

pain come from the disciplines of anesthesiology,

surgery, neurology, and others They have undertaken

extensive training about pain, its causes, and its control,

and as such they represent a major resource for other

health care professionals

At the University of Texas M D Anderson Cancer

Center, we are fortunate to have the availability of all

three above-mentioned components of pain

manage-ment, namely a thoracic team, postoperative pain service,and chronic pain service The management of an individ-ual patient is strongly dependant on the root cause of thepain and whether there was preexisting pain Tolerance is

a phenomenon common to almost all drugs Chronicpain patients can be expected to have much increasedneeds postoperatively Optimum management is facili-tated by early involvement of the chronic pain servicewhen indicated Figure 1-3 shows the organization ofpostoperative pain control and use of pain services at

M D Anderson Cancer Center

Most commonly, patients undergoing thoracic surgeryare opioid naive In our setting they receive an epiduralplaced by a member of the thoracic team or postopera-tive pain service, undergo surgery, and, after activation ofthe epidural analgesia near the end of surgery, have theirpain managed by the postoperative pain service Thepostoperative pain service writes the epidural orders,deals with complications, treats breakthrough pain, and

is proficient in relieving disaggregated pain Each patient

is seen twice per day by a combination of pain servicenurses, rotating residents, and dedicated anesthesiafaculty member, either singly or together Completearound-the-clock coverage is provided, backed up, ifnecessary, by the in-house anesthesiologist on call At allstages the thoracic surgeon is involved and consulted.After the epidural is discontinued, usually at the time ofchest tube removal, the patient’s pain managementreverts completely to the attending thoracic surgeon Thepostoperative pain and chronic pain services are stillavailable in case of need

Thoracotomy Patient

PCEA by PPS Until Epidural Discontinued

IV – PCA by Thoracic Surgery

Opiod

Naive

Successful Good Pain Relief Discharge PPS

Inadequate Poor Pain Relief Other Symptoms

Chronic Pain

on Narcotics Preoperatively

Normal Chronic Pain CPS

Discharge Discharge PPS CPS

Discharge Discharge

FIGURE 1-3 Organization of postoperative thoracic pain management at the University of Texas M D Anderson Cancer Center CPS = chronic

pain service; IV-PCA = intravenous, patient-controlled analgesia; PCEA = patient-controlled epidural analgesia; PPS = postoperative pain service.

Patients with chronic pain, defined variously as being

on > 10 mg or the equivalent of morphine per day, are

equally common at M D Anderson Their management

is similar until discontinuation of the epidural analgesia

At this point the chronic pain service continues their

pain management The chronic pain service is consulted

early in the postoperative period to facilitate changeover

It has been noted that although chronic pain patients

require higher initial and continuing concentrations of

epidural narcotics, their pain relief can still be well

managed by epidural analgesia

If epidural analgesia is impossible, then conventional

intravenous analgesia is employed and initial

manage-ment is by the operating room anesthesiologist After the

patient’s time in the postanesthesia care unit, pain

management falls squarely on the thoracic surgeon Only

if the patient’s pain is impossible to control does the

postoperative pain service become involved Its staff may

be able to suggest different drug regimens or dosages

Chronic pain patients are managed similarly, except that

when intravenous patient-controlled analgesia fails, the

chronic pain service is consulted first

Pre-emptive Analgesia

Pre-emptive analgesia is a subject of endless fascination to

most anesthesiologists It has been defined as the

phenomenon by which analgesia administered prior to a

painful event, such as a thoracotomy, decreases the later

intensity of perceived pain, even after the duration of

action of the initial analgesic.246More recently, the

defini-tion has been widened to include treatment that “prevents

the development of hyperexcitability, even if it takes place

after surgery.”247 It must be distinguished from merely

improved analgesia following earlier administration of

analgesia compared with nonadministration of an

anal-gesic or lesser doses of the same analanal-gesic In a

well-designed study, Doyle and Bowler found that pre-emptive

intravenous morphine, intramuscular diclofenac, and

intercostal nerve blocks with bupivacaine only

demon-strated a decrease in VAS pain scores during a vital

capac-ity breath postoperatively.248No other measure of pain was

different between groups, and the long-term pain was

indistinguishable between groups Early activation of

epidural analgesia might be expected to demonstrate

pre-emptive analgesia, but this expectation was shown to be

incorrect in a study using epidural bupivacaine.249

Pre-emptive analgesia has been suggested with the opioids

fentanyl and morphine in nonthoracic models,250,251raising

the suspicion that higher intraoperative opioid doses may

have a beneficial effect beyond the operative period, but

this has not been demonstrated for thoracotomy patients

It is possible that the high-dose opioid employed for

cardiac surgery in the 1980s and 1990s, using a mediansternotomy approach, led to the oft-repeated observationthat median sternotomy is not as painful as the traditionalthoracotomy With the current fast-track micromanage-ment of cardiac anesthesia, post-sternotomy pain isbecoming a bigger issue for thoracic surgeons It is notclear how an appropriately blinded study could bedesigned to investigate this intriguing possibility

NSAIDs have also been investigated for pre-emptiveeffects but, again, not in thoracic models Using ketoro-lac, Fletcher and colleagues demonstrated improved painrelief in a hip-fracture model at the price of increasedblood loss,252 and Norman and colleagues showed lesspostoperative pain in an ankle fracture model.253Theselective COX-2 inhibitors may have a role to play inpostoperative, pre-emptive, and multimodal analgesia,but so far, proof is lacking

Summary

We have reviewed the provision of postoperative sia for patients who have undergone thoracotomy fromits humble beginnings to the current state of this art—and an art it is, indeed, as the identical drugs and tech-niques in different hands give quite different degrees ofsuccess The provision of pain relief after thoracotomy is

analge-a complicanalge-ated but still rewanalge-arding proposition Newermultimodal analgesia techniques show promise forimproving the overall experience in future patients It islikely that methods to ameliorate or even prevent thedevelopment of chronic post-thoracotomy pain will beelucidated We must all continue to investigate, try differ-ent strategies, improvise, and experiment to ensurecontinued progress in this challenging field

4 El-Baz NM, Faber LP, Jensik RJ Continuous epidural infusion of morphine for treatment of pain after thoracic surgery: a new technique Anesth Analg 1984;63:757–64.

5 Shulman M, Sandler AN, Bradley JW, et al Postthoracotomy pain and pulmonary function following epidural and systemic morphine Anesthesiology 1984;61:569–75.

6 Conacher ID Post-thoracotomy analgesia Anesthesiol Clin North Am 2001;19:611–25.

7 Ready LB Acute postoperative pain In: Miller RD, editor.

Anesthesia 4th ed New York: Churchill Livingstone;

1994 p 2327–44.

8 Richardson J, Sabanathan S, Shah R Post-thoracotomy

spirometric lung function: the effect of analgesia J

Cardiovasc Surg 1999:40:445–56.

9 Scawn NDA, Pennefather SH, Soorae A, et al Ipsilateral

shoulder pain after thoracotomy with epidural analgesia:

the influence of phrenic nerve infiltration with lidocaine.

Anesth Analg 2001;93:260–4.

10 Burgess FW, Anderson DM, Colonna D, et al Ipsilateral

shoulder pain following thoracic surgery Anesthesiology

1993;78:365–8.

11 Tan N, Agnew NM, Scawn ND, et al Suprascapular nerve

block for ipsilateral shoulder pain after thoracotomy with

thoracic epidural analgesia: a double-blind comparison of

0.5% bupivacaine and 0.9% saline Anesth Analg

2002;94:199–202.

12 Nomori H, Horio H, Fuyuno G, Kobayashi R

Non-serraus-sparing antero-axillary thoracotomy with disconnection of

anterior rib cartilage: improvement in postoperative

pulmonary function and pain in comparison to

posterolat-eral thoracotomy Chest 1997;111:572–6.

13 Nomori H, Horio H, Fuyuno G, Kobayashi R Anterior

limited thoracotomy with intrathoracic illumination for

lung cancer Chest 1999;115:874–80.

14 Hazelrigg SR, Landreneau RJ, Boley TM, et al The effect of

muscle-sparing versus standard posterolateral thoracotomy

on pulmonary function, muscle strength, and

postopera-tive pain J Thorac Cardiovasc Surg 1991;101:394–400.

15 Khan IH, McManus KG, McCraith A, McGuigan JA.

Muscle sparing thoracotomy: a biomechanical analysis

confirms preser vation of muscle strength but no

improvement in wound discomfort Eur J Cardiothorac

Surgery 2000;18:656–61.

16 Landreneau RJ, Pigula F, Luketich JD, et al Acute and

chronic morbidity differences between muscle-sparing

and standard lateral thoracotomies J Thorac Cardiovasc

Surg 1996;112:1346–51.

17 Nomori H, Horio H, Naruke T, Suemasu K What is the

advantage of a thoracoscopic lobectomy over a limited

thoracotomy procedure for lung cancer surgery? Ann

Thorac Surg 2001;72:879–84.

18 Mersky H Classification of chronic pain: description of

chronic pain syndromes and definition of pain terms.

Pain 1986;24 Suppl 1:S138–9.

19 d’Amours RH, Riegler FX, Little AG Pathogenesis and

management of persistent postthoracotomy pain Chest

Surg Clin North Am 1998;8:703–22.

20 Dajczman E, Gordon A, Kreisman H, Wolkove N term postthoracotomy pain Chest 1991;99:270–4.

Long-21 Perttunen K, Tasmuth T, Kalso E Chronic pain after thoracic surgery: a follow-up study Acta Anaesthesiol Scand 1999;43:563–7.

22 Katz J, Jackson M, Kavanagh BP, Sandler AN Acute pain after thoracic surger y predicts long-term post- thoracotomy pain Clin J Pain 1996;12:50–5.

23 Nomori H, Horio H, Suemasu K Comparison of term versus long-term epidural analgesia after limited thoracotomy with special reference to pain score, pulmonary function and respiratory muscle strength Surg Today 2001;31:191–5.

short-24 Landreneau RJ, Naunheim K, Magee MJ, Ferson PF Prevalence of chronic pain after pulmonary resection by thoracotomy or video-assisted thoracic surgery J Thorac Cardiovasc Surg 1994;107:1079–86.

25 Peng PWH, Sandler AN A review of the use of fentanyl analgesia in the management of acute pain in adults Anesthesiology 1999;90:576–99.

26 Sandler AN, Baxter AD, Katz J, et al A double-blind, placebo-controlled trial of transdermal fentanyl after abdominal hysterectomy: analgesic, respiratory and phar- macokinetic effects Anesthesiology 1994;81:1169–80.

27 Reinhart DJ, Goldberg ME, Roth JV, et al Transdermal fentanyl system plus IM ketorolac for the treatment of postoperative pain Can J Anaesth 1997;44:377–84.

28 Ashburn MA, Streisand J, Zhang J, et al The sis of fentanyl citrate in humans Anesthesiology 1995;82:1146–53.

iontophore-29 Coluzzi PH, Schwartzberg L, Conroy JD Jr, et al Breakthrough cancer pain: a randomized trial comparing oral transmucosal fentanyl citrate (OTFC ® ) and morphine sulphate immediate release (MSIR ® ) Pain 2001;91:123–30.

30 Lichtor JL, Sevarino FB, Joshi GP, et al The relative potency of oral transmucosal citrate compared with intravenous morphine in the treatment of moderate to severe postoperative pain Anesth Analg 1999;89:732–8.

31 Alexander-Williams JM, Rowbotham DJ Novel routes of opioid administration Br J Anaesth 1998;81:3–7.

32 Azevedo VMS, Lauretti GR, Pereira NL, Reis MP Transdermal ketamine as an adjuvant for postoperative analgesia after abdominal gynecological surgery using lido- caine epidural blockade Anesth Analg 2000;91:1479–82.

33 Woolf CJ, Thompson SW The induction and

mainte-nance of central sensitization is dependent on

N-methyl-D -aspartic acid receptor activation; implications for the treatment of post-injury pain hypersensitivity states Pain 1991;44:293–9.

34 Kawamata T, Omote K, Sonada H, Kawamata M Analgesic mechanisms of ketamine in the presence and absence of peripheral inflammation Anesthesiology 2000;93:520–8.