1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References


This randomized clinical trial compared early outcomes after radiofrequency ablation (RFA) and conventional surgery for varicose veins.


Consecutive patients with symptomatic varicose veins due to isolated great saphenous vein (GSV) incompetence and suitable for RFA were randomized to either RFA or conventional surgery (saphenofemoral disconnection and stripping). Clinical, radiological and patient-based outcomes were recorded at 1 and 5 weeks after intervention.


RFA resulted in successful obliteration of the GSV in all 47 patients. Complete above-knee stripping was unsuccessful in seven of 41 patients. RFA took longer than conventional surgery: median interquartile range 76 (67–84) versus 48 (39–54) min; P < 0·001. Patients returned to their normal activities significantly earlier after RFA (median 3 (2–5) versus 12·5 (4–21) days; P < 0·001). Postoperative pain was significantly less after RFA (median score on visual analogue scale 1·70 (0·50–4·30) versus 4·0 (2·35–6·05); P = 0·001). Patient satisfaction, quality of life improvement and analgesic requirements significantly favoured RFA.


RFA took longer to perform but resulted in a significantly better early outcome than conventional surgery in suitable patients with great saphenous varicose veins. Registration number: ISRCTN29015169 ( Copyright © 2009 British Journal of Surgery Society Ltd. Published by John Wiley & Sons, Ltd.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Lower limb varicose veins are the commonest of all the vascular disorders. They may impair quality of life (QoL)1, 2. Surgery has been shown to correct this impairment3, 4. Over 30 000 operations are performed annually for varicose veins within the National Health Service in the UK5. Thus, despite their benign nature, varicose veins have considerable clinical and socioeconomic impact.

The commonest underlying cause is incompetence of the great saphenous venous system. Conventional surgery to eliminate this abnormal venous reflux involves saphenofemoral disconnection at the saphenofemoral junction (SFJ) and stripping of the great saphenous vein (GSV) above the knee, frequently accompanied by phlebectomies of minor varicosities. Stripping is not always successful and usually no assessment is made as to whether the stripper has indeed passed down the GSV. Unsuccessful stripping is likely to increase the risk of recurrence6, 7. Groin dissection and stripping may cause significant early morbidity owing to bruising, haematoma, pain, cutaneous nerve injury and wound infection8–10, and delay recovery and return to work. These factors often lead to patient dissatisfaction and can result in litigation11.

Minimal access endovenous techniques are becoming popular. They involve ablating the GSV using chemical or thermal energy, thus avoiding vein stripping and groin dissection. This might potentially reduce early postoperative morbidity, hasten recovery and enable an early return to normal activities12–14. The VNUS® Closure® procedure (Covidien (UK) Commercial Ltd, Gosport, UK) is used to obliterate the vein lumen with high-frequency alternating current delivered via an endoluminal catheter with bipolar electrodes, placed percutaneously under duplex ultrasonography12. Several studies have demonstrated that the procedure is safe, with a vein occlusion rate of 90–100 per cent and a patient satisfaction rate of over 90 per cent15–17. The first randomized clinical trial, which included 28 patients in Finland, compared radiofrequency ablation (RFA) with conventional surgery and demonstrated reduced pain, shorter sickness absence and faster return to normal activities in those who underwent RFA18. At 3 years RFA was associated with slightly poorer outcome in terms of recurrent varicosities and reflux, although differences were not statistically significant19. A second multi-institutional randomized study involved 80 legs and showed similar early outcomes20. Overall outcomes for RFA at 2 years were at least equivalent to those of conventional surgery21.

The present randomized clinical trial was conducted to evaluate clinical, radiological and patient-based outcomes after RFA and conventional surgery in a selected population. The long-term follow-up is designed to be completed 5 years after intervention. This paper records the short-term results from this trial.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The trial was conducted in the specialist vascular unit of a university hospital. The local Research Ethics Committee approved the study. Consecutive patients aged between 18 and 70 years who presented to a one-stop venous clinic with symptomatic varicose veins (Clinical Etiologic Anatomic Pathophysiologic (CEAP) clinical class 2–6) underwent clinical assessment. Those who satisfied the selection criteria (Table1) underwent duplex ultrasonography. Patients identified as suitable for RFA were consented and allocated to receive either RFA or conventional surgery by a web-based randomization method that used two stratification variables—age (50 years or less and over 50 years) and sex (male or female)—that were judged most likely to influence the outcome in the two groups. The CEAP clinical class22, Michael's classification group23, Total Clinical Severity Score (TCSS)24, Venous Disability Score (VDS) and American Society of Anesthesiologists fitness grade were recorded22–24. A baseline estimate of QoL was obtained through validated self-administered disease-specific tools: the Aberdeen Varicose Vein Questionnaire (AVVQ) and VEINES-QoL/Sym questionnaire (V-Q/Sym Q)1, 25, 26.

Table 1. Inclusion and exclusion criteria
  1. GSV, great saphenous vein; RFA, radiofrequency ablation; QoL, quality of life.

Inclusion criteria
 Age between 18 and 70 years, either sex, elective
 GSV reflux (primary or recurrent) on duplex imaging and
   requiring surgery
 Duplex scan confirmed suitability for RFA (see exclusion
 Patient fit for a general anaesthetic
 Physical condition allowing ambulation after the procedure
 Patient able to give informed consent
 Requirement for intervention agreed between patient and
 Availability of patients for all follow-up visits
Exclusion criteria
 Varicose veins without GSV incompetence on duplex
 Associated small saphenous or deep venous
   incompetence on duplex imaging
 Tortuous GSV above the knee felt to be unsuitable for
 GSV diameter < 3 mm or > 12 mm in the supine position
 Thrombus in the GSV
 Patients with a pacemaker or internal defibrillator
 Concomitant peripheral arterial disease (ankle : brachial
   pressure index < 0·9)
 Unable to complete QoL questionnaires owing to poor
   English language skills


All patients received a single subcutaneous prophylactic dose of low molecular weight heparin at the time of intervention. All operations were performed under general anaesthesia. In both groups phlebectomy hooks were used for simultaneous avulsion of varicosities that had been marked before operation.

Radiofrequency ablation

RFA was performed by surgeons with sufficient experience27. Before skin preparation, the duplex scan was used to map the course of the GSV in the thigh and mark the vein access site at knee level (SonoSite TITAN® L38 5–10 MHz; SonoSite, Bothell, Washington, USA). With the patient in the reverse Trendelenburg position, the GSV was accessed percutaneously by the Seldinger technique and a 6- or 8-Fr FAST-CATH Haemostasis Introducer (Daig Corporation, Minnetonka, Minnesota, USA) inserted. The VNUS® Closure® PLUS intravascular catheter with bipolar electrodes (6 or 8 Fr depending on vein diameter) was introduced through the sheath and positioned in the GSV with its tip just below the entry of the superficial epigastric vein. The tissues overlying the GSV were infiltrated under duplex guidance with saline (500 ml 0·9 per cent saline containing 20 ml 1 per cent lidocaine with 1 : 200 000 epinephrine) to achieve vein compression. The thigh was wrapped with an Esmarch bandage from the knee to the groin with the leg elevated and the patient placed in the Trendelenburg position.

The position of the tip of the catheter was confirmed and its proximal end connected to the radiofrequency generator. The electrodes were unsheathed and wall contact was tested by measuring the impedance. The target temperature was set at 85 °C. The ablation was commenced just distal to the entry of superficial epigastric vein and the catheter was pulled back at the rate of 1·5–2 cm per min for the first 3 cm and 2–3 cm per min for the remainder of the procedure. Heparinized saline (5000 units heparin in 500 ml 0·9 per cent saline) was infused continuously via the central lumen of the catheter to prevent thrombus formation on the electrodes. The Esmarch bandage was then removed and the sheath withdrawn to treat the lowest segment of the vein.

A duplex scan was performed on completion of the procedure. Successful ablation was indicated by one of three flow patterns: absence of spontaneous or augmented flow, augmented flow with thickened vein wall, or spontaneous trickle flow with thickened vein wall and vein calibre 50 per cent of less of pretreatment size. Any retreatment of inadequately ablated vein segments was performed immediately.

Conventional surgery

Conventional surgery was performed with the patient in the Trendelenburg position by a trained surgeon. Through a skin-crease groin incision the SFJ was exposed and tributaries of the GSV were ligated and divided. The GSV was ligated and divided close to the SFJ (high ligation). A perforate–invagination (PIN) stripper was passed down the open distal end of the vein to emerge at knee level and the skin incised at this point to retrieve the stripper. The vein was secured to the upper end of the stripper, which was retrieved by pulling it down to knee level and out of the exit wound, thus stripping the vein. The groin wound was infiltrated with 10 ml 0·5 per cent bupivacaine and closed in two layers with absorbable sutures.

Wound closure and dressing

In both groups, phlebectomy wounds were closed with Steri-Strips (3M Health Care, St Paul, Minnesota, USA) and covered with water-resistant adhesive dressings. A crepe bandage was applied to the whole leg and replaced by appropriately sized graduated compression stockings (thigh length) after 24 h, to be worn for 2 weeks.

Postintervention advice

Patients were encouraged to walk as frequently as possible (a few hundred metres three times a day at least), to avoid prolonged standing and to sit with the feet elevated. A discharge pack containing 30 tablets of co-codamol (8/500 mg) and four of diclofenac (75 mg) was provided for pain relief and patients were encouraged to use them as necessary. Patients were advised to return to normal activities when they felt able to do so, and to avoid driving until they could perform an emergency stop without any discomfort. No fixed duration of sick absence was advised and patients were encouraged to return to work when they felt sufficiently well and comfortable, taking the nature of their job into consideration.


Patients were seen at the end of first and fifth weeks after the intervention. A duplex scan was carried out by an independent sonographer at the first follow-up visit. The primary outcome was the time taken to return to normal household activities. Successful treatment was defined on duplex imaging as the complete absence of any blood flow from 3 cm below SFJ along the length of the treated vein segments (RFA) or complete disappearance of the GSV (conventional surgery). Secondary outcomes included intraoperative complications, duration of the procedure, postoperative morbidity (pain, analgesic requirements, sensory abnormalities, wound problems, phlebitis, skin burns, pigmentation), time to return to driving, and patient satisfaction and QoL at second follow-up. Pain was evaluated using a 10-cm unmarked visual analogue scale (VAS) and patient-reported grading of severity (none, very mild, mild, moderate, severe, very severe). Patient satisfaction was assessed using a 10-cm unmarked VAS and a grading system (A to D, with A rated best) based on their willingness to recommend the procedure to others, or to undergo the same procedure on the opposite leg, if required.

Statistical analysis

The power calculation was based on the results from published studies18, 20, 28. The primary outcome measure was the time taken to return to full normal household activities. Following conventional surgery, about 40 per cent of patients return to normal activities within a week of surgery. It was hypothesized that RFA could bring about a 75 per cent improvement in this outcome and such an improvement would have clinical significance. To detect such a difference with 80 per cent power and a significance level of 5 per cent, a sample size of 84 patients, 42 in each arm, was required. This was increased by 10 per cent to account for any potential loss to follow-up.

χ2 test or Fisher's exact test was used for comparison of proportions, and t test for continuous data. Minitab release 13.32 (Minitab, State College, Pennsylvania, USA) was used for analyses. QoL data were analysed by means of SPSS® for Windows® version 10.0 (SPSS, Chicago, Illinois, USA) using recommended scoring methods and syntax files.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

A total of 128 patients were assessed, of whom 93 were randomized. Eighty-eight patients underwent the selected intervention as a day-case procedure; five were not treated (Fig.1). Forty-seven patients had RFA (13 men, 34 women) and 41 had conventional surgery (14 men, 27 women), with a median (interquartile range (i.q.r.)) age of 47 (38–58) and 45 (37–53) years respectively. There was no crossover of patients between the treatment arms after randomization and before treatment. No conversion of RFA to conventional surgery was required for failure to achieve ablation. All patients achieved planned first and second follow-up at a median (i.q.r.) of 6 (5–7) and 37 (34–42) days respectively. The groups were evenly matched (Table2).

thumbnail image

Figure 1. CONSORT diagram for the trial. *Patient preference for conventional surgery (three), patient decided against intervention (three)

Download figure to PowerPoint

Table 2. Type, distribution and severity of disease, and fitness grade
 Radiofrequency ablation (n = 47)Conventional surgery (n = 41)
  • CEAP, Clinical Etiologic Anatomic Pathophysiologic; TCSS, Total Clinical Severity Score; VDS, Venous Disability Score; ASA grade, American Society of Anesthesiologists fitness grade.

  • *

    One patient C4 and TCSS 4;

  • one patient C6 and TCSS 12.

Michael's classification
 Group 22720
 Group 31719
CEAP clinical class
 ≥ 41*1
ASA grade


There were no significant differences between the groups in terms of anaesthetic induction or recovery, or discharge from the day surgery unit. The total theatre time (between entry into and exit from the theatre suite) was significantly longer for RFA than for conventional surgery (Table3). The total procedure time between marking the course of the GSV under duplex guidance in RFA or antiseptic preparation of the operative field in conventional surgery to application of compression bandage in both instances was similarly longer for RFA, indicating that the observed time difference was not due to extraneous factors.

Table 3. Main outcomes after radiofrequency ablation and conventional surgery for great saphenous varicose veins
 Radiofrequency ablation*Conventional surgery*95 per cent c.i. for difference between groupsP
  • *

    Values are median (interquartile range).

  • c.i., Confidence interval; VAS, visual analogue scale.

  • t test.

Theatre time (min)82 (73–91)55 (48–63)22·40, 33·41< 0·001
Procedure time (min)76 (67–84)48 (39–54)24·35, 35·32< 0·001
Pain in first week (VAS score)1·70 (0·50–4·30)4·00 (2·35–6·05)−2·75, −0·790·001
Duration of analgesia (days)2 (0–7)10 (5–14)−8·61, −1·140·011
Return to normal activity (days)3 (2–5)12·5 (4–21)−11·61, −4·23< 0·001
Return to driving (days)4 (2–6)7 (5–12·5)−6·82, −2·19< 0·001
Return to work (days)10 (4–13)18·5 (11–28)−12·95, −2·210·006
Satisfaction (VAS score)10 (8·4–10)8·5 (7·5–10)0·15, 1·440·016

The median (i.q.r.) pull-back time and pull-back speed of the RFA catheter were 14·0 (12·6–15·7) min and 2·22 (1·96–2·55) cm/min respectively. The number of stab wounds above and below the knee as well as their timing (whether performed during or after the main part of the operation) were comparable between the two groups and had no impact on the time differences observed between the procedures. Duplex imaging after RFA revealed successful ablation in all but two procedures. These veins showed segmental spontaneous brisk flow and were retreated immediately with success. Procedural difficulties encountered with RFA included unsuccessful tumescent infiltration owing to venous spasm and poor visualization on the duplex scan (two procedures), GSV perforation by the ablation catheter close to the venepuncture site (two) and close to the SFJ (one), and catheter tip thrombus that required withdrawal of the catheter and flushing (nine). Skin burns were not encountered in the trial.

The only intraoperative problem encountered with conventional surgery was incomplete stripping in seven patients. In five patients the upper two-thirds of the GSV was stripped successfully, whereas the upper third alone was stripped in another instance. In one patient (with recurrence after 15 years) it was impossible to strip any part of the GSV owing to inability to pass the PIN stripper down the vein. Reverse stripping was successful on one occasion, a retriever was used to complete stripping in four patients and a sizeable anterior accessory saphenous vein was stripped on one occasion.

Follow-up duplex imaging

Duplex imaging at first follow-up revealed a patent untreated GSV stump with antegrade flow across the SFJ in 45 of 47 patients after RFA, with a median (i.q.r.) stump length of 9·9 (7·9–12·3) mm. Two patients had complete absence of flow across the SFJ. The GSV was visualized with no demonstrable flow in all patients who underwent RFA. In the conventional surgery group, all patients had total obliteration of flow across the SFJ. Of the seven patients with incomplete GSV stripping, one demonstrated brisk reflux and two demonstrated no flow; the others showed trickle retrograde flow. All deep veins were patent in both groups.

Clinical outcomes

Minor complications

A significantly higher rate of cutaneous sensory abnormalities was observed after conventional surgery (Table4). These were most frequently observed along the medial aspect of the leg in both groups. The commonest paraesthesia experienced was a tingling sensation. Groin wound problems noted after conventional surgery included mild inflammation (three patients), serous wound discharge (two), haematoma (one) and wound breakdown (one), all of which resolved spontaneously. Clinically evident haematomas in the thigh and leg were slightly more common after conventional surgery but did not differ significantly between the groups. Five patients developed a non-tender palpable GSV with overlying pigmentation after RFA that showed progressive resolution by the second follow-up.

Table 4. Sensory abnormalities after treatment
 Radiofrequency ablation (n = 47)Conventional surgery (n = 41)P*
  • *

    χ2 test.

Numbness/reduced sensation
 1-week follow-up9200·003
 5-week follow-up7190·003
 1-week follow-up5110·049
 5-week follow-up650·936
Postoperative pain

Postoperative pain and analgesic requirements were considerably less following RFA (Table3, Fig. 2). Pain scores beyond the first week similarly favoured RFA. Seven patients experienced no pain after ablation (recorded 0 on pain VAS) and ten did not require any analgesia. In contrast, all patients experienced pain after conventional surgery and all except one required postoperative analgesia.

thumbnail image

Figure 2. Severity grading of pain in the first week after intervention. RFA, radiofrequency ablation

Download figure to PowerPoint

Total Clinical Severity Score and Venous Disability Score

Before the procedure 42 of 47 patients in the RFA group and 39 of 41 in the conventional surgery group had a TCSS of 2 or less; 46 and 37 respectively had a VDS of 1 or less (Table2). Despite this, more than three-quarters in each group had an improved TCSS and VDS after surgery. Two patients, one in each group, had worsening of VDS by 1 point after treatment. Thus both procedures proved effective in improving symptoms in a significant proportion of patients.

Time to return to activities, driving and work

Patients returned to their full level of normal household activities, to driving and to work significantly more quickly following RFA than after conventional surgery (Table3). After a conventional operation, 15 of 41 patients returned to normal activities within a week. The power calculation was based on a 75 per cent expected improvement in this outcome after RFA, but the observed improvement was over 100 per cent, with 35 of 47 returning to usual activities within 5 days of RFA (P < 0·001).

Similar results were observed with driving (66 patients were regular drivers), with 34 of 38 patients able to resume regular driving within a week of RFA compared with 15 of 28 after conventional surgery (P < 0·001). Sixty-six of 88 trial participants were in employment with no significant difference between the groups, the majority being in ‘on-the-feet’ employment. Two patients after RFA and six after a conventional procedure required minor adjustments at work for a limited period when they resumed work.

Patient satisfaction

This was significantly better after RFA (Table3). Twenty-seven of 47 patients were completely satisfied (score 10 on satisfaction VAS), compared with 11 of 41 who had conventional surgery (P = 0·004). Significantly more patients in the RFA group were willing to recommend the operation to others (46 of 47 versus 32 of 41; P = 0·005) (Table5). The most common reason for dissatisfaction in both groups was postoperative morbidity (pain, discomfort, restricted mobility). No patient in either group expressed complete dissatisfaction with the procedure (score 0 on satisfaction VAS or satisfaction grade D).

Table 5. Grading system for recording patient satisfaction (second follow-up visit)
 Radiofrequency ablation (n = 47)Conventional surgery (n = 41)
A—willing to recommend the4632
   operation to others  
B—willing to consider the14
   operation for the opposite  
   leg, if required, but not sure  
   about recommending it to  
C—not sure whether would05
   consider the operation for  
   the opposite leg if required  
D—will never have the00
   operation again or never  
   recommend to anyone else  
Quality of life

Analysis of AVVQ data revealed significantly improved QoL for the whole group after intervention. The improvement was greater after RFA than after conventional surgery but the difference did not reach statistical significance (mean improvement in QoL score −9·12 versus −8·24; 95 per cent c.i. for difference −3·64 to 1·89; P = 0·532).

V-Q/Sym Q analysis showed a significant improvement in both symptoms and QoL for the whole group after intervention. There was a greater improvement in symptom score following RFA than after conventional surgery (mean improvement 12·62 versus 9·94; 95 per cent c.i. for difference −1·65 to 7·01; P = 0·220). QoL improvement was significantly greater after RFA (mean improvement 12·80 versus 7·83; 95 per cent c.i. 0·80 to 9·14; P = 0·020).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

RFA is a minimally invasive procedure. Its potential early benefits, by avoiding groin dissection and GSV stripping, have been confirmed by the findings from this trial. Patients experienced less pain and recovered more quickly, which was reflected in better patient satisfaction and QoL outcomes following RFA. Previous trials have mirrored these favourable findings early after RFA18, 20. One of these trials was performed partly during the surgeons' learning curve18. The other trial had some methodological flaws: there was considerable variability between the two groups and between centres with regard to the stripping technique and type of anaesthesia used20. Improved QoL scores for RFA at 72 h and 1 week may have been influenced by recall bias. An attempt was made to address these issues in the present trial.

The majority of patients in this trial had primary symptomatic uncomplicated varicose veins (CEAP clinical class C2–3). Randomization with stratification ensured an even distribution between the groups in terms of age and sex, and disease severity profile. Surgeons with sufficient experience performed all the procedures. Postintervention advice was also standardized. It was, however, difficult to blind the patients, operator or the independent observer to the intervention.

RFA took significantly longer to perform than conventional surgery. This was due to tumescent infiltration and detailed duplex scanning before and after treatment. However, the procedure time was still considerably quicker than the duration of RFA from an earlier trial18. Duplex imaging is not used routinely to assess the completeness of stripping after conventional surgery. The new VNUS® ClosureFAST catheter produces ‘segmental ablation’ resulting in a significantly faster treatment time. RFA is routinely performed under tumescent infiltration alone and without an Esmarch bandage in many centres, and the authors have now adopted this approach. These factors would also reduce the duration of RFA treatment.

RFA involves considerably less tissue dissection and trauma. Patients who had RFA experienced considerably less pain, required less analgesia and were able to return to their activities more quickly. This improved confidence, lifestyle, social contacts, cosmesis, QoL and satisfaction in the early days after the procedure. Driving is a good index of recovery from a procedure such as varicose vein surgery. Most trial participants resumed driving significantly more quickly after RFA. Similarly, patients who underwent RFA returned to work at least a week earlier than those who had conventional surgery. This was also noted in an earlier trial18.

Patient satisfaction is often recorded as a willingness to recommend the intervention to a friend or family member17. It is more appropriate to use a grading system or a more objective evaluation29. The most common reason for dissatisfaction was postoperative morbidity, illustrating the impact of adverse sequelae. QoL was significantly better after both interventions (the study was not powered specifically to detect any difference in this outcome). Scores were, however, consistently higher after RFA. Previous studies have shown that this difference persisted even at 2 years21.

Current evidence based on randomized trials consistently demonstrates significant early benefits after RFA in suitable patients with varicose veins. Five-year outcomes from large cohort studies appear comparable to those after conventional surgery30. It is important, however, to remember that conventional surgery, unlike RFA, may be performed on most patients irrespective of the configuration or distribution of varicosities, and that stripping of the GSV is effective in reducing long-term recurrence and risk of reoperation31, 32. The conventional procedure also improves QoL and symptoms in most patients, and differs from RFA only in the extent of the improvement and in short-term morbidity. Long-term results from this and other similar studies may throw more light on the consequences of leaving intact the superficial epigastric vein and other tributaries at the SFJ. They may also reveal the cost-effectiveness of the procedure, which will depend on late outcomes. This should help establish the precise role of RFA in the treatment of varicose veins.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

VNUS Medical Technologies (San Jose, California, USA) funded some of the Closure® PLUS radiofrequency ablation catheters used in the trial. They were not involved in the running of the trial, data collection, interpretation or analyses. The authors take full academic responsibility for the content of this manuscript, and declare no personal conflict of interest.

The authors thank the Institute of Health and Society, Newcastle University, UK, for their assistance in setting up the web-based randomization process and in the use of QoL assessment tools, the Day Surgery Unit and Outpatients Department at Freeman Hospital, where the follow-up clinics were held, and consultants and colleagues who contributed to the study.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References