Radioactive seed localization of nonpalpable breast lesions is better than wire localization

BACKGROUND: The current study sought to validate radioactive seed localization (RSL) as an alternative to wire localization (WL) to facilitate the operative excision of nonpalpable breast lesions. METHODS: One hundred consecutive patients underwent preoperative WL and the next 100 RSL. Margins were considered negative if > or =2 mm from in situ and invasive disease. RESULTS: RSL resulted in 100% retrieval of the seeds and lesions. Sixty-eight percent of patients underwent RSL at least 1 day before surgery. RSL resulted in a 35% relative improvement in the rate of negative margins in the first specimen (P = 0.01) and a 62% relative improvement in the rate of reoperation for positive margins (P = 0.01). The sentinel lymph node (SLN) identification rate was 100% in both groups. CONCLUSIONS: RSL is effective and safe, and this procedure significantly improved the rate of negative margins in the first specimen and the rate of reoperation for positive margins compared to WL. We highly favor RSL over WL.     

A Randomized Prospective Comparison of Patient‐Assessed Satisfaction and Clinical Outcomes with Radioactive Seed Localization versus Wire Localization

Radioactive seed localization (RSL) has emerged as an alternative to wire localization (WL) in patients with nonpalpable breast cancer. Few studies have prospectively evaluated patient satisfaction and outcomes with RSL. We report the results of a randomized trial comparing RSL to WL in our community hospital. We prospectively enrolled 135 patients with nonpalpable breast cancer between 2011 and 2014. Patients were randomized to RSL or WL. Patients rated the pain and the convenience of the localization on a 5‐point Likert scale. Characteristics and outcomes were compared between groups. Of 135 patients enrolled, 10 were excluded (benign pathology, palpable cancer, mastectomy, and previous ipsilateral cancer) resulting in 125 patients. Seventy patients (56%) were randomized to RSL and 55 (44%) to WL. Fewer patients in the RSL group reported moderate to severe pain during the localization procedure compared to the WL group (12% versus 26%, respectively, p = 0.058). The overall convenience of the procedure was rated as very good to excellent in 85% of RSL patients compared to 44% of WL patients (p < 0.0001). There was no difference between the volume of the main specimen (p = 0.67), volume of the first surgery (p = 0.67), or rate of positive margins (p = 0.53) between groups. RSL resulted in less severe pain and higher convenience compared to WL, with comparable excision volume and positive margin rates. High patient satisfaction with RSL provides another incentive for surgeons to strongly consider RSL as an alternative to WL.     

Randomized prospective evaluation of a novel technique for biopsy or lumpectomy of nonpalpable breast lesions: radioactive seed versus wire localization

Background: Standard wire localization (WL) and excision of nonpalpable breast lesions has several shortcomings.Methods: Ninety-seven women with nonpalpable breast lesions were prospectively randomized to radioactive seed localization (RSL) or WL. For RSL, a titanium seed containing ¹²⁵I was placed at the site of the lesion by using radiographical guidance. The surgeon used a handheld gamma detector to locate and excise the seed and lesion.Results: Both techniques resulted in 100% retrieval of the lesions. Fewer RSL patients required resection of additional margins than WL patients (26% vs. 57%, respectively; P = .02). There were no significant differences in mean times for operative excision (5.4 vs. 6.1 minutes) or radiographical localization (13.9 vs. 13.2 minutes). There were also no significant differences in the subjective ease of the procedures as rated by surgeons, radiologists, and patients. All WLs were carried out on the same day as the excision, whereas RSL was performed up to 5 days before the operative procedure.Conclusions: RSL is as effective as WL for the excision of nonpalpable breast lesions and reduces the incidence of pathologically involved margins of excision. RSL also reduces scheduling conflicts and may allow elimination of intraoperative specimen mammography. RSL is an attractive alternative to WL.     

A Multicentered, Randomized, Controlled Trial Comparing Radioguided Seed Localization to Standard Wire Localization for Nonpalpable, Invasive and in situ Breast Carcinomas

Background

Studies suggest radioguided seed localization (RSL) yields fewer positive margins than wire-guided localization (WL). The goal of this study is to determine whether RSL is superior to WL.

Methods

Women with confirmed invasive or ductal carcinoma in situ (DCIS) undergoing localization and breast conserving surgery were enrolled. Outcomes measured include positive margin and reoperation rates, specimen weight, operative and localization times, and surgeon and radiologist ranking of procedural difficulty.

Results

Randomization was centralized, concealed, and stratified by surgeon with 153 patients in the WL group and 152 in RSL group. Localizations were performed using either ultrasound (70%) or mammographic guidance (30%). Pathology was either DCIS (18%) or invasive carcinoma (82%). Procedures were performed at 3 sites, by 7 surgeons. Only difference found for patient and tumor characteristics was more multifocal disease in RSL group. Using intention-to-treat analysis, there were no differences in positive margins rates for RSL (10.5%) and WL (11.8%), (P = .99) or for positive or close margins (<1 mm) (RSL 19% and WL 22%; P = .61). Mean operative time (minutes) was shorter for RSL (RSL 19.4 vs WL 22.2; P < .001). Specimen volume, weight, reoperation and localization times were similar. Surgeons ranked the seed technique as easier (P = .008), while radiologists ranked them similarly. Patient’s pain rankings during wire localization were higher (P = .038).

Conclusions

In contrast to other trials positive margin and reoperation rates were similar for RSL and WL. However, for RSL operative times were shorter, and the technique was preferred by surgeons, making it an acceptable method for localization.     

Radioactive Seed Localization Compared to Wire Localization in Breast-Conserving Surgery: Initial 6-Month Experience

Background

Wire localization (WL) of nonpalpable breast cancers on the day of surgery is uncomfortable for patients and impacts operating room efficiency. Radioactive seed localization (RSL) before the day of surgery avoids these disadvantages. In this study we compare outcomes of our initial 6-month experience with RSL to those with WL in the preceding 6 months.

Methods

Lumpectomies for invasive or intraductal cancers localized with a single ¹²⁵iodine seed (January–June 2012) were compared with those using 1 wire (July–December 2011). Surgeons and radiologists did not change. Positive and close margins were defined as tumor on ink and tumor ≤1 mm from ink, respectively. Demographic and clinical characteristics and outcomes were compared between RSL and WL patients.

Results

There were 431 RSL and 256 WL lumpectomies performed. Clinicopathologic characteristics did not differ between groups. Most seeds (90 %) were placed before the day of surgery. Positive margins were present in 7.7 % of RSL versus 5.5 % of WL patients, and 16.9 % of RSL versus 19.9 % of WL had close margins (p = 0.38). The median operative time was longer for lumpectomy and sentinel lymph node biopsy (SLNB) in the RSL group (55 vs. 48 min, p < 0.0001). There was no significant difference in the volume of tissue excised between groups.

Conclusions

In the first 6 months of RSL, operative scheduling was simplified, while rates of positive and close margins were similar to those seen after many years of experience with WL. Operative time was slightly longer for RSL lumpectomy and SLNB; we anticipate this will decrease with experience.     

Adequacy of invasive and in situ breast carcinoma margins in radioactive seed and wire-guided localization lumpectomies

Image-guided preoperative localizations help surgeons to completely resect nonpalpable breast cancers. The objective of this study is to compare the adequacy of specimen margins for both invasive breast cancer (IBC) and ductal carcinoma in situ (DCIS) after radioactive seed localization (RSL) vs wire-guided localization (WGL). We retrospectively reviewed 600 cases at a single Canadian academic center from January 2014 to September 2017, comparing surgical margins, re-excisions and reoperations, localization accuracy and major complications (migration, accidental deployment, vasovagal reaction), as well as operative duration between RSL and WGL cases. IBC margins were positive in 7% of RSL and 6% of WGL cases (P = .57). Tumor size (P = .039) and association with DCIS (P = .036) predicted positive margins in invasive carcinoma. DCIS margins were positive in 6% and 8%, and close (≤2 mm) in 37% and 36% of cases (P = .45) for RSL and RSL cases respectively. The presence of extensive intraductal component predicted positive DCIS margins (P < .0001). There was no significant difference between intraoperative re-excisions (P = .54), localization accuracy (P = .34), and operation duration (P = .81). Reoperation for lumpectomies and mastectomies was marginally higher for WGL than RSL (P = .049). There were 11 (4%) WGL and no RSL complications (P = .03). Overall, positive margins for IBC, close or positive margins for DCIS, intraoperative re-excision, localization accuracy, and operation duration were similar between RSL and WGL. The reoperation rate was higher in WGL than RSL, which may reflect practice changes over time. RSL was safer than WGL with lower complication rates.     

Positive margins after breast-conserving therapy: localization technique or tumor biology?

Background

The relative contributions of patient and tumor factors versus radiologic localization technique to the rates of inadequate margins of excision in breast-conserving therapy have not been defined.

Methods

Patients undergoing breast-conserving therapy were studied. Margins less than 2 mm from tumor were considered inadequate.

Results

Of 539 patients, 31% were guided by palpation and 69% were guided by preoperative radiologic localization. The palpation-guidance patients had larger tumors (P < .0001) and more nodal metastases (P = .0005). The rates of inadequate margins were 10% for palpation-guided patients and 11% for radiologic-localization patients (P = .53). The 3-year rates of local recurrence were .7% for palpation-guided patients and 1.8% for radiologic-guided patients (P = .5).

Conclusions

Patient, tumor, and intraoperative pathologic factors, not just localization device shortcomings, produce inadequate margins of excision in breast-conserving therapy. A reasonable expected rate of inadequate margins owing to patient and tumor factors is 10%. Quality improvement for margin management must focus on intraoperative assessment of margins, especially for patients with identified risk factors, in addition to improving localization technique.     

Radioactive seed vs wire localization for nonpalpable breast lesions: A single institution review

The aim of this study was to compare wire localization (WL) and radioactive seed localization (RSL) for nonpalpable breast lesions with regard to margin status, re‐excision rate, procedure length, and complications related to localization. A retrospective review of the electronic health records at a single institution was performed. There was no difference in re‐excision rate, margin positivity, volume of tissue removed, and complication rate for RSL vs WL (P = 0.9934, P = 0.9934, P = 0.6645, and P = 0.4716 respectively). The only difference was a longer OR time, RSL = 104.408 minutes vs WL = 82.386 minutes. (P = 0.0163). RSL and WL are comparable techniques for localization of nonpalpable breast lesions.     

The introduction of radioactive seed localisation improves the oncological outcome of image guided breast conservation surgery

IntroductionRadioactive seed localisation (RSL) has become increasingly popular for localisation of non-palpable breast tumours. This is largely due to advantages it offers in terms of practicality and convenience when compared to guide wire localisation (WL). This institute switched from using WL to RSL in September 2014. The primary aim was to assess whether this change improved the accuracy of excision with regards to inadequate margin rates and weight of excision specimens. The secondary aim was to establish whether there is a “learning curve” associated with RSL technique.MethodsRetrospective data collection was performed for 333 consecutive cases of unifocal non-palpable invasive breast cancers undergoing excision with WL or RSL. An inadequate margin was defined as tumour <1 mm from an inked radial margin. Patient demographics, tumour characteristics and clinical outcomes were compared between WL and RSL cases.Results100 WL and 233 RSL cases were included. Patient demographics and tumour characteristics were similar for both groups. Inadequate margin rates were 18% with WL and 8.6% with RSL (p = 0.013). Median specimen weights were 33.3 g with WL and 28.7 g with RSL (p = 0.014). Subdividing the RSL group into the first 100 cases performed (RSL1) and the subsequent 133 cases (RSL2), inadequate margin rates were 13.0% and 5.3% respectively (p = 0.037). Mean specimen weights were similar.ConclusionSwitching from WL to RSL results in a significant reduction in both inadequate margin rates and specimen weights. A procedure-specific learning curve is present on first implementation of RSL and following this, inadequate margin rates are further reduced.     

Radioactive Seed Localization Breast Biopsy and Lumpectomy: Can Specimen Radiographs Be Eliminated?

Background: Wire localization (WL) is the current standard for surgical diagnosis of nonpalpable breast lesions. Many disadvantages inherent to WL are solved with radioactive seed localization (RSL). This trial investigated the ability of RSL to reduce the need for specimen radiographs and operating room delays associated with WL.Methods: A total of 134 women were entered onto an institutional review board–approved study. RSL was performed by placing a titanium seed containing .29 to 20 mCi of iodine-125 to within 1 cm of the suggestive breast lesion. The surgeon used a handheld gamma detector to locate and excise the iodine-125 seed and the lesion.Results: Specimen radiographs were eliminated in 98 (79%) of 124 patients. Surgical seed retrieval was 100% in 124 patients. No seed migration occurred after correct radiographical placement. A total of 26 (21%) of 124 patients required a specimen radiograph; 22 (85%) of these 26 were performed for microcalcifications.Conclusions: After surgical removal, RSL can eliminate specimen radiographs when the radiologist accurately places the seed and the pathologist grossly identifies the lesion. If small microcalcifications are noted before surgery, then specimen radiographs may be necessary. RSL reduced requirements for specimen radiographs, decreased OR time, improved incision placement, and improved resections to clear margins.     

MR Imaging Features of Triple‐Negative Breast Cancers

Triple‐negative (TN) breast cancers, which are associated with a more aggressive clinical course and poorer prognosis, often present with benign imaging features on mammography and ultrasound. The purpose of this study was to compare the magnetic resonance imaging features of TN breast cancers with estrogen (ER) and progesterone (PR) positive, human epidermal growth factor receptor (HER2) negative cancers. Retrospective review identified 140 patients with TN breast cancer who underwent a preoperative breast MRI between 2003 and 2008. Comparison was made to 181 patients with ER+/PR+/HER2? cancer. Breast MRIs were independently reviewed by two radiologists blinded to the pathology. Discrepancies were resolved by a third radiologist. TN cancers presented with a larger tumor size (p = 0.002), higher histologic grade (<0.001), and were more likely to be unifocal (p = 0.018) compared with ER+/PR+/HER2? tumors. MRI features associated with TN tumors included mass enhancement (p = 0.026), areas of intratumoral high T2 signal intensity (p < 0.001), lobulated shape (p < 0.001), rim enhancement (p < 0.001), and smooth margins (p = 0.005). Among the TN tumors with marked necrosis, 26% showed a large central acellular zone of necrosis.     

Radioactive seed localization of breast lesions: an adequate localization method without seed migration

Preoperative localization is important to optimize the surgical treatment of breast lesions, especially in nonpalpable lesions. Radioactive seed localization (RSL) using iodine-125 is a relatively new approach. To provide accurate guidance to surgery, it is important that the seeds do not migrate after placement. The aim of this study was to assess short-term and long-term seed migration after RSL of breast lesions. In 45 patients, 48 RSL procedures were performed under ultrasound or stereotactic guidance. In the first 12 patients, the lesion was localized with two markers: an iodine-125 seed and a reference marker. In 33 patients, 36 RSL procedures were performed using a single iodine-125 seed. All patients received control mammograms after seed placement and prior to surgery. In the patients with two markers, migration was defined as the difference in the largest distance between the markers observed in the mammograms. For single-marked lesions, migration was assessed by comparing distances between anatomical landmarks in the mammograms. RSL was successful in all patients. Seeds were in-situ for 59.5 days on average (3-136 days). The detection rate during surgery was 100%. Overall, an average seed migration of 0.9 mm (standard deviation 1.0 mm) was observed. Neither differences in lesion type, nor days in situ, type of surgery or radiologic localization method were found to have impact on seed migration. RSL is an accurate preoperative localization method for breast lesions with negligible seed migration, independent of time in-situ.     

Long-Term Results of Breast Conservation in Chinese Women with Breast Cancer

Abstract:  Between July 1989 and December 2002, 172 women with Stage I/II breast cancer were treated by breast conservation therapy (BCT). All underwent quadrantectomy and axillary node clearance. Minimum follow-up was 5 years and 79 (52%) were followed for >10 years. At 5 years, local relapse-free and overall survival rates were 98.3% and 98.3%. The 10-year rates were 95% and 94%, respectively. The 10-year local recurrence rate was higher in patients with involved margins (33.3% versus 2.7%, p = 0.0272). Furthermore 10-year death rates in margin positive patients were higher (18.2% versus 2.5%, p = 0.0486). Excellent or good cosmetic results were achieved in 54%. BCT is a reasonable option for early stage breast cancer in Chinese women but margin status is the most important determinant of local recurrence. Negative margins are required for optimal local control and minimization of distant metastasis.     

Efficacy of localization of non-palpable,invasive breast cancer: Wire localization vs. Iodine-125 seed: A historical comparison

BackgroundTo achieve radical tumour excision in non-palpable breast cancer accurate tumour localization is essential. Historically, this has been achieved by wire localization (WL). Recently, new techniques like Iodine-125 seeds (IL) have become available. The aim of this study was to determine whether the results in case of IL are better than when WL is used.MethodA consecutive series of patients operated on because of non-palpable breast cancer was analysed. Between January 2007 and April 2011 WL was method of choice. From April 2011 till January 2013 IL was introduced and became the standard procedure.Primary outcomes weresuccess rate of (peri-) operative localization and radicality of the excision. Secondary outcomes were duration of operation and volumes of the excised tissue specimen.ResultsA total of 236 women were included; 149 women underwent WL and 87 women IL. The overall success rate for WL was 93.3%. In 10 patients (6.7%) WL failed because of peri-operative dislocation of the wire compared to only one failure to locate the seed in de postoperative specimen in the IL-group (1.1%, p = 0.098).Radical resection was obtained in 126/149 (84.6%) of the WL-patients, in comparison with 81/87 (93.1%) in the IL-group (p = 0.054).Median operation time and volume of the specimen was not significantly different between both groups.ConclusionLocalization of non-palpable breast cancer using Iodine-125 seeds is, at least, as good as the standard wire localization procedure.     

The Yale University Experience of Early-Stage Invasive Lobular Carcinoma (ILC) and Invasive Ductal Carcinoma (IDC) Treated with Breast Conservation Treatment (BCT): Analysis of Clinical-Pathologic Features, Long-Term Outcomes, and Molecular Expression of COX-2, Bcl-2, and p53 as a Function of Histology

Abstract:  To evaluate our experience of the clinical-pathologic features and outcomes of early-stage Invasive Lobular (ILC) versus Invasive Ductal (IDC) carcinoma treated with breast conservation treatment (BCT). 142 ILC and 1,760 IDC patients were treated with BCT at our institution. All patients underwent breast conserving surgery and radiation therapy (median total dose: 64 Gy). Clinical-pathologic and outcome parameters were analyzed to detect differences between the two cohorts. In addition, COX-2, Bcl-2, and p-53 expression was analyzed from our existing tissue micro-array database. Median follow-up was 6.8 years. A higher percentage of ILC patients presented at >40 years of age (94% ILC versus 89% IDC, p = 0.0353) and had more mammographically occult tumors (p < 0.002). There were no significant differences in T stage, nodal status, family history, final margin, ER/PR/HER-2 status or triple negative tumors (all p-values >0.05). From the immuno-histochemical analysis, expression of p53, COX-2, and Bcl-2 did not differ significantly (all p-values >0.05) between the two cohorts. At 10 years, there was no difference in breast relapse (20% versus 13%, p = 0.25), distant relapse (26% versus 20%, p = 0.28), cause-specific survival (72% versus 84%, p = 0.09) and OS (68% versus 78%, p = 0.08). Patients with ILC had higher contralateral breast relapses (26% versus 12%, p = 0.0006). Patients with early-stage ILC have comparable outcomes to IDC when treated with BCT. Because of the higher risk of contralateral breast cancers for ILC patients, careful evaluation of the contralateral breast will be important in the follow-up of these patients. Future investigations of chemo-preventive strategies to decrease contralateral breast cancers are warranted.     

Health‐related quality of life may improve after transplantation in pancreas–kidney recipients

Pancreas–kidney transplantation (PKT) may significantly improve quality of life (HRQOL) in patients with type 1 diabetes. We have assessed the changes felt by PKT patients, using the Gastrointestinal Quality of Life Index (GIQLI) and EuroQol‐5D questionnaires. Patients were asked to compare how their HRQOL had changed from pre‐transplantation to the last visit. The 60 men and 66 women enrolled had a mean follow‐up of five yr; 84.1% with both grafts, 15.9% with one graft functioning. In all domains of EuroQol‐5D scores improved after PKT, as well as the visual analogue scale health state (from 38% to 84%, p < 0.001; effect size 3.34). In GIQLI, physical function was felt better after PKT than before (14.83 ± 3.86 vs. 7.86 ± 4.43, p < 0.001; effect size 1.68); the same was observed for psychological status, social function, and GI complaints. Concerning the burden of medical treatment, the score significantly improved (from 1.31 to 3.63, p < 0.001, effect size 2.02). The rate of unemployed patients decreased after PKT (from 50.8% to 36.5%, p < 0.001). Multivariate analysis showed that having only one functioning graft was associated with worse HRQOL scores (B = ?5.157, p = 0.015). In conclusion, for all assessed domains, patients reported a significant improvement in HRQOL after PKT. Maintenance of the two grafts functioning predicted higher improvement of HRQOL scores.     

Localizing High-Risk Lesions for Excisional Breast Biopsy: A Comparison Between Radioactive Seed Localization and Wire Localization

Background

Improved resolution and utilization of screening breast imaging has increased identification of nonpalpable high-risk lesions (HRL) and subsequent excisional breast biopsies (EBBs). Wire localization (WL), used most commonly for EBBs, may have shortcomings, including wire displacement, patient discomfort, limitations with incision planning and scheduling logistics. Radioactive seed localization (RSL) may overcome these drawbacks. The purpose of this study was to compare WL and RSL for EBBs for HRLs.

Methods

All single-site EBBs for HRL performed by four breast surgeons were retrospectively reviewed over two consecutive 1-year periods. Patients with cancer on percutaneous core biopsy (CB) were excluded. Clinicopathologic information, operative time, targeted lesion retrieval rate, and upstage rate were collected.

Results

A total of 324 EBBs for HRL were performed: 196 using WL and 128 using RSL. CB pathology was atypical hyperplasia in 56 % of WLs and 62 % of RSLs. The remaining pathologies were radial scar, papilloma, atypical papilloma or lobular carcinoma in situ. Mean age was 54 years. OR time was 27 ± 8 min for WL and 27 ± 7 min for RSL (p = 0.9). Upstage rate was 6 and 5 % for WLs and RSLs, respectively (p = 0.5). Targeted lesions were retrieved in 98 % of WL and 99 % of RSL (p = 0.5). SV was 37.2 ± 32.8 cm³ and 25.7 ± 22.3 cm³ for WL and RSL, respectively (p = 0.001).

Conclusions

RSL is comparable to WL for EBB of HRLs with similar OR times and upstage rates. SV is significantly decreased with RSL and may translate into improved cosmetic outcomes without sacrificing the diagnostic accuracy of the EBB.     

Prognostic Evaluation of Breast Cancer Patients with Evident Bone Marrow Metastasis

In this study, we aimed to evaluate the clinicopathologic characteristics and prognosis of breast cancer (BC) patients with symptomatic bone marrow metastasis (BMM). Fifty‐four BC patients, including patients with and without BMM, were evaluated retrospectively. In particular, the clinicopathologic features and survival of the patients with BMM (n = 27) were assessed and compared with the patients without BMM. All of the patients with BMM also had osseous metastases, and bone was the first site for distant recurrence in the majority of patients in the study group. Anemia was the most frequent symptom at presentation. The median time to BMM was 36.1 months (range 1.6–70.5 months, 95% CI). HER2(+) patients developed BMM earlier than HER2(?) patients (3.2 versus 38.3 months, 95% CI; p = 0.05). Patients with advanced disease at the time of initial BC diagnosis developed BMM earlier than patients with early disease (p = 0.04). Time to development of BMM was significantly shorter in tumors with perinodal infiltration (p = 0.001) and multicentric focus (p = 0.025). Median survival time after the diagnosis of apparent BMM was 6.43 months. Survival after BMM diagnosis in patients with grade III tumors was significantly shorter than in patients with grade I–II tumors (1.43 versus 5.36 months, 95% CI; p < 0.001). Systemic therapy after BMM diagnosis significantly prolonged survival (17.3 versus 0.93 months, 95% CI; p < 0.001). Hormone receptor‐positive, high‐grade, advanced‐stage tumors at the time of initial BC diagnosis were more common in patients with BMM. Invasive lobular histology was also more frequent in patients with BMM. In conclusion, the presence of hormone receptor‐positive, multicentric, grade III, advanced‐stage tumors may be important risk factors for the development of evident BMM in BC patients. Systemic single‐agent chemotherapy can prolong survival in these patients. However, multicenter analyses are required to verify these findings.     

Thoracic epidural analgesia started after cardiopulmonary bypassAdrenergic, cardiovascular and respiratory sequelae

The effects of thoracic epidural analgesia started after cardiopulmonary bypass were studied on the subsequent adrenergic, cardiovascular and respiratory responses. Sixteen cardiac surgical patients received either a standardised general anaesthetic (control group) or a standardised general anaesthetic and thoracic epidural analgesia (epidural group). The epidural catheter was sited before surgery and heparinisation. Following discontinuation of cardiopulmonary bypass, patients in the epidural group were given 15 ml bupivacaine 0.5% down the catheter followed by an infusion of bupivacaine 0.375% at 5–8 ml.h⁻¹ after surgery. The control group received an intravenous morphine infusion on completion of surgery. The adrenergic response was assessed by measuring arterial catecholamine concentrations. Respiratory function was determined by spirometry, peak expiratory flow and arterial partial pressure of oxygen while breathing air. Pain scores were also obtained. After cardiopulmonary bypass the increases in catecholamine concentrations were effectively inhibited in the epidural group for the remainder of the study (p < 0.05). Postoperative respiratory function was less impaired in the epidural group, with higher forced expiratory volume in 1 s, forced vital capacity and peak expiratory flow (p < 0.05). Pain scores were also significantly lower in the epidural group (p < 0.05). There were no significant differences in cardiovascular parameters.