The thiazoladinediones drugs for type 2 diabetics improve insulin sensitivity, increase the number of adipocytes and reduce the size of adipocytes which I think lead to the idea that more smaller adipocytes = healthy fat.
Here's a fairy recent paper showing that an increasing ratio of leptin to adiponectin is correlated to insulin resistance and more leptin is produce by big adipocytes.
Obesity is the dominant cause of insulin resistance. In adult humans it is characterised by a combination of adipocyte hypertrophy and, to a lesser extent, adipocyte hyperplasia. As hypertrophic adipocytes secrete more leptin and less adiponectin, the plasma leptin:adiponectin ratio (LAR) has been proposed as a potentially useful measure of insulin resistance and vascular risk. We sought to assess the usefulness of the LAR as a measure of insulin resistance in non-diabetic white adults.
Leptin and adiponectin levels were measured in 2,097 non-diabetic individuals from the Ely and European Group for the Study of Insulin Resistance (EGIR) Relationship between Insulin Sensitivity and Cardiovascular Risk (RISC) study cohorts. LAR was compared with fasting insulin and HOMA-derived insulin sensitivity (HOMA-S) in all individuals and with the insulin sensitivity index (M/I) from hyperinsulinaemic–euglycaemic clamp studies in 1,226 EGIR RISC participants.
The LAR was highly correlated with HOMA-S in men (r = −0.58, p = 4.5 × 10−33 and r = −0.65, p = 1.1 × 10−66 within the Ely and EGIR RISC study cohorts, respectively) and in women (r = −0.51, p = 2.8 × 10−36 and r = −0.61, p = 2.5 × 10−73). The LAR was also strongly correlated with the clamp M/I value (r = −0.52, p = 4.5 × 10−38 and r = −0.47, p = 6.6 × 10−40 in men and women, respectively), similar to correlations between HOMA-S and the M/I value.
The leptin:adiponectin ratio is a useful measure of insulin resistance in non-diabetic white adults. These data highlight the central role of adipocyte dysfunction in the pathogenesis of insulin resistance. Given that variations between fasting and postprandial leptin and adiponectin levels tend to be small, the leptin to adiponectin ratio might also have potential value in assessing insulin sensitivity in the non-fasted state.