STARLING PRESSURES IN THE HUMAN ARM AND THEIR ALTERATION IN POSTMASTECTOMY EDEMA

1. Surgery and radiotherapy to axillary lymph nodes during breast canc er treatment is often followed, commonly years later, by chronic postm astectomy oedema (PMO). PMO is considered a 'high protein' oedema due to reduced axillary lymph drainage. Since oedema formation also depend s on fluid input (capillary filtration), we studied the Starling press ures in the affected and contralateral arm. Colloid osmotic pressure w as measured in patient serum (pi(p)) and interstitial fluid (pi(i)). S ubcutis fluid was collected from PMO arms by both wick and aspiration methods, and from the control arm by the wick method only. Interstitia l hydraulic pressure (P-i) was measured by the wick-in-needle method. 2. Oedema pi(i) was 19.2 +/- 4.1. cmH(2)O (n=13, wick) to 16.3 +/- 4.4 cmH(2)O (n = 41, aspirate; difference not significant; means +/- S.D. throughout). This was significantly lower than pi(i) in the control a rm (21.4 +/- 3.8 cmH(2)O, n = 14, P < 0.01, analysis of variance). Als o, there was a negative correlation between oedema pi(i) and the perce ntage increase in arm volume (correlation coefficient r = -0.35, P < 0 .05) in contrast to conventional expectation. 3. Oedema P-i (1.9 +/- 2 .0 cmH(2)O, n = 28) exceeded the subatmospheric control P-i (-2.8 +/- 3.0 cmH(2)O; P < 0.01). Venous and arterial pressures were normal but pi(p) was subnormal (31.1 +/- 2.7 cmH(2)O, n = 47). 4. Net pressure op posing capillary blood pressure, P-o, was calculated as P-i + sigma (p i(p) - pi(i)) for a reflection coefficient, sigma, of 0.90-0.99. P-o i n the control arm, 6.2-8.5 cmH(2)O, was less than antecubital venous p ressure, 10.5 cmH(2)O. This provides no support for the traditional vi ew that venous capillaries in the arm are in a state of continuous flu id reabsorption. 5. In the PMO arm, P-o was raised to 15.0 +/- 4.6 cmH (2)O (n = 28, P < 0.01, ANOVA) and correlated positively with increase in arm volume (r = 0.41, P < 0.05). A rise in P-o will reduce filtrat ion rate if capillary pressure is unaltered, which should raise inters titial protein concentration and pi(i) - whereas pi(i) actually decrea sed. Possible explanations include a rise in capillary pressure. The p athophysiology of postmastectomy oedema evidently involves additional mechanisms besides lymphatic damage.