Relative Eosinophilia (Thorn Test) as a Bioassay to Judge the Clinical Relevance of Cortisol Values during Severe Stress

Albertus Beishuizen and István Vermes

Departments of Internal Medicine and Clinical Chemistry Medical Spectrum Twente, Hospital Group Enschede, the Netherlands 7500 KA

We read with interest the recent editorial by D. H. P. Streeten (1) in JCEM. In his discussion of the paper by Abdu et al. (2), Streeten indicated that the low dose ACTH test cannot conclusively exclude the presence of mild or incomplete deficiencies in the functional capacity of the hypothalamic-pituitary-adrenal (HPA) system. He addressed the impairment of ACTH response to severe stress (3) and the lack of reliable testing of the HPA axis during prolonged stress conditions. In contrast to the mostly increased cortisol levels during severe stress (4), we reported discrepant low ACTH levels in severely ill trauma and septic patients (5). Streeten states that we still have to rely on clinical judgement to determine the adequacy of the measured cortisol levels in patients under stressful conditions, as in intensive care units.

We speculate that the classical Thorn test (6) can help our educated guess in cases of relative adrenal insufficiency. Therefore, we would like to revive the neglected observation by Thorn et al. (6) fifty years ago, when they noted the amount of circulating eosinophils as a biological indicator for adrenocortical function. We wondered if monitoring eosinophils with electronic cell counters, with their superior accuracy and precision, could reveal a decline in adrenocortical function more quickly and with more sensitivity than the manual counting at the time of Thorn’s publication (6). Using electronic cell counters, we measured the percentage of eosinophils in 612 consecutive patients admitted to the intensive care unit (7). Use of corticosteroids or drugs known to influence the HPA axis and a history of adrenal insufficiency led to exclusion. When eosinophilia, defined as more than 3%, was present (in 40 of 570 assessable patients), a standard dose (250 µg) and a low-dose (1 µg) ACTH stimulation test (SST) were performed. It appeared that the low-dose SST was abnormal in 10 of 40 (25%) patients, in contrast to 2 cases in which the standard SST was abnormal (defined as a 30-min post-stimulation cortisol concentration <550 nmol/L). Corticosteroid treatment was given to all 8 patients, in whom a strong clinical suspicion for relative adrenal insufficiency was apparent. An illustrative case is shown in Fig. 1Go. Full recovery was seen in 7 cases, suggesting the presence of adrenal failure. From these preliminary results we think that the low-dose SST might be superior to the standard SST in detecting relative adrenal insufficiency in critically ill patients. Relative eosinophilia offers a very useful bioassay to assess adrenocortical function longitudinally and should be considered a warning signal for insufficient adrenocortical function. Given the absence of optimal biochemical tests to detect (partial) HPA failure during severe stress situations, relative eosinophilia reflects the functional shortage of glucocorticoids.



View larger version (25K):
[in this window]
[in a new window]
 
Figure 1. Percentage of eosinophils (—{blacksquare}—), basal cortisol level (—•—), and mean arterial pressure (—{blacktriangleup}—) during the stay of a patient suffering from septic shock and multiple organ failure. Synacthen stimulation tests (SST) with 250 µg and 1 µg ACTH were performed on days 10 and 11, respectively. The arrows indicate the increment in serum cortisol concentration after the SST. The bar indicates the duration of hydrocortisone treatment (100 mg every 6 h).

 
Accordingly, we suggest that the original Thorn test is still a useful bioassay to assess adrenocortical functionality, in particular because the present electronic cell counters now allow sensitive and reproducible enumeration of eosinophils in blood samples.

The study by Bertelloni et al. (1) in the December, 1998 issue of The Journal of Clinical Endocrinology and Metabolism evaluated lumbar spine bone mineral density (BMD) using dual energy x-ray absorptiometry (DXA) in 21 young men with histories of constitutionally-delayed puberty (mean age 21.8 yr) and 12 control subjects (mean age 19.3 yr). Thirteen of the men with histories of delayed puberty had received androgen therapy. The authors reported that spinal BMD was significantly lower in the men with histories of delayed puberty than in control subjects (1.101 ± 0.134 g/cm2 vs. 1.222 ± 0.091 g/cm2; P < 0.009). These findings confirmed our earlier reports that men with histories of constitutionally delayed puberty have decreased radial and spinal BMD compared with controls (2, 3). Men with histories of delayed puberty are often shorter than normal (4, 5). Because standard DXA measurements are influenced by bone size (so that smaller bones have lower apparent BMD), the authors also analyzed their data after correcting for bone size (volumetric BMD or vBMD). The authors reported that vBMD was similar among the men with histories of pubertal delay and controls. Based on this finding, they concluded that our previous finding of decreased BMD in men with histories of pubertal delay "may be solely the consequence of uncritical use of DEXA."

The authors are correct that alterations in bone size affect standard DXA (or areal BMD) results. Thus, it is reasonable to hypothesize that decreases in BMD in men with histories of pubertal delay could be due to decreases in bone size and thus an artifact of areal BMD measurements. Because of this, we performed a multivariate analysis that included body mass index in our 1992 study and found that differences in BMD between the men with histories of delayed puberty and controls persisted after adjusting for this factor. Moreover, we compared radial width and spinal bone area among the two groups, and they did not differ significantly. Later in 1992, Carter et al. (6) reported a method to correct spinal BMD values for differences in bone size (i.e. bone mineral apparent density or BMAD). The same group also reported a method for correcting radial BMD measurements for projection artifacts (7). We have now reanalyzed our original data using these formulas for determining BMAD.

As indicated in Table 1Go, the magnitude of the decrease in BMD in men with a history of constitutionally-delayed puberty (approximately a 10% decrement) is similar whether areal BMD or BMAD are used to express the data, and the differences remain highly statistically significant. Thus, we stand by our original assertion that BMD is decreased in adult men with histories of pubertal delay and that this finding is not due to a measurement artifact caused by smaller bone size. The reasons for differences in the findings of Bertelloni et al. and our study are not obvious, but they may be related to the fact that most of the subjects in the study by Bertelloni et al. had received androgen therapy during adolescence or to the fact that their control men were only 19 yr old and may not have yet reached peak bone mass.


View this table:
[in this window]
[in a new window]
 
Table 1. BMD and BMAD of the radius and spine in men with histories of delayed puberty and normal men

 
Finally, Bertelloni et al. wrote that we and others recommended that boys with delayed puberty receive early treatment "to ensure more optimal BMD in adulthood" (1). In fact, we were extremely careful not to make such a claim, and we wrote "Although we found that delayed puberty leads to diminished peak bone mineral density in men, we do not know whether early androgen supplementation would normalize the peak bone mineral density of boys with constitutionally delayed puberty and protect them from osteoporosis later in life" (2). We feel that it is important to clarify this point.

Footnotes

Address correspondence to: István Vermes, Department of Clinical Chemistry, Medical Spectrum Twente, Hospital Group, P.O. Box 50000, Enschede, The Netherlands 7500 KA.

Address correspondence to: Joel Stephen Finkelstein, Endocrinology Unit, Massachusetts General Hospital, Bulfinch 327, 55 Fruit Street, Boston, Massachusetts 02114.

Received March 19, 1999.

References

  1. Bertelloni S, Baroncelli GI, Ferdeghini M, Perri G, Saggese G. 1998 Normal volumetric bone mineral density and bone turnover in young men with histories of constitutional delay of puberty. J Clin Endocrinol Metab. 83:4280–4283.[Abstract/Free Full Text]
  2. Finkelstein JS, Neer RM, Biller BMK, Crawford JD, Klibanski A. 1992 Osteopenia in adult men with a history of delayed puberty. N Engl J Med. 326:600–604.[Abstract]
  3. Finkelstein JS, Klibanski A, Neer RM. 1996 A longitudinal evaluation of bone mineral density in adult men with histories of delayed puberty. J Clin Endocrinol Metab. 81:1152–1155.[Abstract]
  4. Crowne EC, Shalet SM, Wallace WBH, Eminson DM, Price DA. 1990 Final height in boys with untreated constitutional delay of puberty. Clin Endocrinol (Oxf). 65:1109–1112.
  5. Rosenfield RL. 1990 Diagnosis and management of delayed puberty. J Clin Endocrinol Metab. 70:559–562.[Medline]
  6. Carter DR, Bouxsein ML, Marcus R. 1992 New approaches for interpreting projected bone densitometry data. J Bone Miner Res. 7:137–145.[Medline]
  7. Katzman DK, Bachrach LK, Carter DR, Marcus R. 1991 Clinical and anthropometric correlates of bone mineral acquisition in healthy adolescent girls. J Clin Endocrinol Metab. 73:1332–1339.[Abstract]