Difficulties with defining lymphoedema after axillary dissection for breast cancer

Assessment of axillary node status remains one of the single most important prognostic indicators in breast cancer, and may influence choice of adjuvant therapies. Until recently, Axillary Node Dissection (AND) has been the standard surgical technique to assess these nodes.A serious side effect of AND is lymphoedema (LO). LO is a build up of lymphatic fluid in the arm caused by damage to arm lymphatic drainage when axillary lymph nodes are removed.12 With moderate or severe lymphoedema, the affected arm can be painful, tired and heavy.6,1,12,13 The excess lymphatic fluid acts as a culture medium and the disrupted lymph flow prevents a normal immune response making the arm1 more susceptible to infection.Patients are advised to take particular care of their affected limb and to seek medical treatment promptly if infection develops to try to minimise LO risk.11 Long term LO is accompanied by subcutaneous and lymphatic fibrosis.6The current literature around assessment of LO is confusing. There are several different measurement techniques in use, and consensus on definition of LO, particularly with arm circumference measures is poor. The incidence of LO is also changing over time as surgery and treatment techniques change. Due to these two factors the reported incidence ranges from 2-56%.1This wide variation makes it difficult to compare studies and to know how a particular locality measures up to the published literature. It is important for both patients and surgeons to know the local risk for developing lymphoedema after AND. There is only one study to date in New Zealand to report local incidence of LO and it was retrospective study based on postal questionnaires.20Several risk factors for developing LO have been previously determined, such as: treating the axilla with axillary radiation after AND, which causes tissue fibrosis, and chronic lymphoedema by constricting lymphatic channels.12The primary aims of this study were to ascertain the local prevalence of lymphoedema after AND and to compare the relationship between the objective measures of LO and the subjective reporting of arm complications. Given the difficulty with LO definition, we decided to examine simple methods of assessment, in a subset, to determine which was best.Secondary aims of the study were: to see if there were any local risk factors that affect the rate of LO; to measure the morbidity associated with LO using QoL questionnaire and to establish an easy and convenient way to detect LO patients in surgical clinic after ANDMethods The participant population comprised consecutive women attending a breast cancer follow-up clinic at Waikato Hospitals Breast Care Centre or at a local surgeons private rooms. Woman who met eligibility criteria were recruited when an interviewer was available. Exclusion criteria included bilateral surgery; pre-existing lymphoedema prior to AND; less than 3 months after surgery; and surgery not carried out in the Waikato. Those who consented underwent an interview and arm assessment. The assessments were conducted by four trained staff members. The assessment consisted of a questionnaire followed by an examination in which the patients arm circumferences were measured. The questions were designed to find out risk factors for lymphoedema (arm work, post op breast/axillary wound infection or seroma collection, air flight travel, intravenous cannula to operated arm, arm infection or cellulitis and arm injury). Pathological and treatment details were obtained from medical records. Quality of Life (QoL) questionnaire16 included; activities requiring reaching overhead, driving car for > 15min, pulling shirt overhead, combing hair, doing up a back fastening bra, pushing a supermarket trolley with both hands, making a bed, zipping a back fastening dress, wiping down a table top, doing usual sporting activities (total of 10 questions and patients were asked to circle 0-4; 0=unknown, 1=no difficulty, 2=some problem, 3=very difficult, 4=unable to perform). Scores were summed for analysis. Arm measurement was conducted using a tape measure. All of the participants had their arms measured 15cm above and 10 cm below the olecranon, and around the hand. The hand measurement was conducted by asking the participant to make a fist with the thumb on the outside of the fist, then measuring the circumference of the widest point, which is the base of the thumb and mid metacarpal. To measure arm circumference for each point, minimal pressure with the tape was used to avoid compressing the arm soft tissues. 73 patients towards the end of study also had another 7 measurements starting from base of middle finger proximally every 10cm apart up to arm. A tape measure was used to mark points at 10cm intervals from middle finger on both hands with most distal point marked at 60cm from base of middle finger. Arm volumes were calculated using Casley-Smith method for calculation of volume of a truncated cone.17 Subjective questions were asked to ascertain which women had experienced ongoing problems with LO. They were asked to circle on a scale of 1-4 (1=no problems, 2=a little, 3=quite a bit and 4=very much), if since surgery they had experienced: arm swelling; heaviness; or tightness. For the purposes of analysis they were grouped as under: Group 2+ that included those who circled 2, 3 or 4 for any of the questions; and Group 3+that circled 3 or 4 for any of the questions. Statistical analysisData collected during interviews was then entered into a Microsoft Access database. We used multivariate logistic regression method for analysis of LO risk factors. A binary variable (lymphoedema Y/N) was the dependent variable. Statistical analysis for the Activities of daily living questionnaires was performed using GraphPad Prism (version 5.0). Continuous variables were compared using the Mann Whitney test. A p-value of <0.05 was retained as statistically significant. Arm dominance correctionA correction was made for arm dominance using a factor of 1.4% for the forearm measurement, 1.2% for the hand measurement and 0% for the upper arm measurement. These figures were devised by using a subset of 105 women who said they had never had arm swelling and did not have any detectable swelling on measurement (at the 7.5% threshold). These women most probably do not have lymphoedema. The measurements of the dominant and non-dominant arms of these women were compared and the difference was found as detailed above. These figures are comparable with those found in the literature. Kannus et al (1995) 14 found that the difference was 0.7% in the upper arm and 1.2% in the forearm. Another study 15 found the difference was 2.5%. Both of these studies had different population groups than the women in this study {healthy young controls} so the correcting factor determined in our own population was considered to be more accurate, and is what was used in the analysis. Results 193 women with complete data were analysed. Patient demographics are shown in Table 1. Table 1. Patient demographics Variables Mean 00b1 SD Age Weight Height BMI Time since surgery 61 00b1 11 yrs 74 00b1 17 kg 157 00b1 30 cm 28 00b1 6 Kg/m2 56.42 00b1 37.48 (3-183 months) Lymphoedema in our study was defined as 22657.5% increase in any circumference in the operated arm compared to non operated (control) arm after arm dominance correction. A 7.5% increase in circumference at all points is comparable to 15.5% increase in volume using formula: Area = C 00b2/4pi. A 7.5% increase at just one point, therefore indicates a much smaller increase in arm volume. A 10% increase in arm circumference corresponds to a 20% increase in arm volume, if this occurs at all points and considerably less if just at one point. A 226520% increase in arm volume is defined as moderate lymphoedema by an International Consensus Group.19 We also calculated the prevalence of LO using several different thresholds used in the literature (Table 2). Table 2. Prevalence of lymphoedema using subjective and objective criteria Criteria used to define lymphoedema Percentage (number) Subjective 2+ score (a little arm swelling or more) Subjective 3+score (quite a bit of arm swelling or very much) 41.9% (81) 10.8% (21) Any measurement 226510% increase in operated arm All 3 measurements 226510% in the operated arm Any measurement 22657.5% increase in operated arm All 3 measurements 22657.5 in the operated arm Any measurements 22655% in the operated arm All measurements 22655% in the operated arm Any measurement 22652cm increase in operated arm 12.9% (25) 0.0% (0) 23.3% (45) 1.0% (2) 40.9% (79) 2.1% (4) 25.3% (49) Volume 226520% increase in operated arm Volume 226515% increase in operated arm Volume 226510% increase in operated arm 8.2% (6/73) 9.6% (7/73) 19.1% (14/73) In a subset of 73 patients out of 193, serial measurements were done along with 3 basic measurements and we calculated % increase in volume in the operated arm compared to the other arm as control by using Casley-Smith method.17 In this subset, the prevalence of LO was 8.2%, 9.6% and 19.1% using volume increase cut offs of 20%, 15% and 10% compared to non operated arm respectively. International consensus guidelines for management of LO19 define the LO in terms of volume increase as Mild < 20%, Moderate 20-40% and Severe >40% volume increase in operated arm. None of our women developed severe LO. Prevalence of LO using subjective methodi.e. subjective arm swelling at any time since surgery, was 10.8% (3+ score) and 41.9% (2+ score). Prevalence using 2265 2cm increase in any circumference in operated arm was 25.3%. We calculated the sensitivity and specificity of the circumference and subjective methods by comparing 226520% increase in the volume as gold standard in the subgroup of 73 patients (Table 3). Table 3. Sensitivity and Specificity of Subjective and Objective methods: Different methods to detect lymphoedema Sensitivity Specificity Any measurement 22657.5% inc Op arm Any measurement 226510% inc Op arm Any measurement 22652cm inc Op arm Subjective 3+ (quite a bit and very much) 83% 66% 66% 67% 81% 89% 80% 93% Based on that, a 7.5% increase in circumference in the operated arm compared to the non operated arm has a high sensitivity to detect LO but still will call 19% of women with mild or no LO as having LO. Concordance between subjective and objective measurements is shown in Table 4. Table 4. Concordance between the subjective and objective measurements Variables Y/N 22657.5% increase in any measurement (Total 45) Number (Percentage) Arm swelling score 3+ NO 31 (69%) (quite a bit and very much) YES 14 (31%) Arm swelling score 2+ NO 11 (24%) (a little arm swelling or more) YES 34 (75%) Variables Y/N 226510 % increase in any measurement (total 25) Number (Percentage) Arm swelling score 3+ NO 13 (52%) (quite a bit and very much) YES 12 (48%) Arm swelling score 2+ NO 2 (8%) (a little arm swelling or more) YES 23 (92%) Variables Y/N Arm swelling 2+ Arm swelling 3+ 22657.5% increase in any measurement NO 47 7 YES 34(42%) 14 (67%) Total 81 21