Measuring stress in animals is important in many ecological, zoological and veterinary research settings. A common method is to measure plasma levels of glucocorticoid hormones (cortisol, corticosterone, hereafter CORT). Over the past decade, an alternative method has become widely popular: assessing leucocyte profiles; in other words, the heterophil‐to‐lymphocyte ratio (H/L) in birds and reptiles or the neutrophil‐to‐lymphocyte ratio (N/L) in mammals, amphibians and fish. Recent studies have shown that although both indices reliably change after acute stress, they are not always correlated at baseline, and thus are not interchangeable. This lack of correlation has led to some confusion about which of these two measures is “better”; in other words, more biologically relevant and a truer readout of stress.
In this review, we first document the dramatic expansion of the use of leucocyte profiles to assess stress over the past decade. We then review published evidence that may explain the discrepancies between leucocyte profiles and plasma CORT. We argue that these discrepancies stem from the nature of the stress reaction and the role of corticosterone (or cortisol) versus leucocytes during the stress response.
The differing roles of the two responses translate to differences in the timing of each, which is important for researchers studying either acute or chronic stress. Although plasma glucocorticoids rise within minutes of the onset of acute stress and return to baseline within 1 or 2 hr after the stressor passes, H/L or N/L ratios remain low for at least 30–60 min or sometimes more.
The temporal nature of these metrics differs also in the context of long‐term chronic stress. The glucocorticoid response wanes in repeatedly stressed animals, or those exposed to chronic environmental stress, which can often (but not always) lead to low baseline levels and more typically, an attenuated acute stress response. Meanwhile, H/L or N/L ratios appear to remain high under the same conditions.
Understanding the temporal differences in the responses should help researchers decide which metric is the better method in their particular study, or whether both would be useful for elucidating different elements of the stress response.