D A Butt1, E B Sochett. 1. Division of Endocrinology, Hospital for Sick Children, Toronto, Ontario, Canada.
Abstract
OBJECTIVES: The majority of short statured children referred for serum GH testing prove to be GH sufficient. The purpose of our study was to evaluate urinary growth hormone (uGH) as a screening test for GH sufficiency. PATIENTS: We studied (i) short statured children previously diagnosed as GH sufficient (n = 44) or GH deficient (n = 41) (peak serum GH > or = 8 micrograms/l or < 8 micrograms/l, respectively); (ii) short children undergoing serum GH stimulation tests (n = 23, test group); (iii) normal statured children (n = 45, control group). DESIGN: Three separate overnight urine collections were obtained in all groups. GH injections in GH deficient subjects were discontinued 4 days prior to urine collection. MEASUREMENTS: uGH concentrations were measured using a chemiluminescence immunoassay. Overnight uGH was expressed in several ways (overnight excretion and overnight excretion corrected for body surface area, time and creatinine). Receiver operator curves (ROC) were constructed from the data obtained in the GH sufficient and deficient subjects. Sensitivity and specificity were then determined for various urinary cut-offs. These cutoffs were validated in turn in the test group by comparison of the predicted with the observed GH status. RESULTS: The GH deficient group had the lowest GH output with respect to overnight uGH, overnight uGH/m2, overnight uGH/h and overnight uGH/creatinine when compared with the GH sufficient and control groups (P = 0.0001). Overnight uGH/m2 data gave the greatest area under the ROC curve. At 100% specificity (no GH deficient subjects), it had the highest sensitivity, 63.6% (49.2-78.0% CI) at a cut-off of 2.3 ng/m2 (63.6% of GH sufficient subjects had uGH levels > 2.3 ng/m2). When this and other cut-offs were applied to the test group, we found consistency between the observed and predicted numbers of GH sufficient and deficient subjects. CONCLUSIONS: We conclude that urinary GH is a useful test for the diagnosis of GH sufficiency as defined by serum criteria and can be used to reduce significantly the number of serum stimulation tests.
OBJECTIVES: The majority of short statured children referred for serum GH testing prove to be GH sufficient. The purpose of our study was to evaluate urinary growth hormone (uGH) as a screening test for GH sufficiency. PATIENTS: We studied (i) short statured children previously diagnosed as GH sufficient (n = 44) or GH deficient (n = 41) (peak serum GH > or = 8 micrograms/l or < 8 micrograms/l, respectively); (ii) short children undergoing serum GH stimulation tests (n = 23, test group); (iii) normal statured children (n = 45, control group). DESIGN: Three separate overnight urine collections were obtained in all groups. GH injections in GH deficient subjects were discontinued 4 days prior to urine collection. MEASUREMENTS: uGH concentrations were measured using a chemiluminescence immunoassay. Overnight uGH was expressed in several ways (overnight excretion and overnight excretion corrected for body surface area, time and creatinine). Receiver operator curves (ROC) were constructed from the data obtained in the GH sufficient and deficient subjects. Sensitivity and specificity were then determined for various urinary cut-offs. These cutoffs were validated in turn in the test group by comparison of the predicted with the observed GH status. RESULTS: The GH deficient group had the lowest GH output with respect to overnight uGH, overnight uGH/m2, overnight uGH/h and overnight uGH/creatinine when compared with the GH sufficient and control groups (P = 0.0001). Overnight uGH/m2 data gave the greatest area under the ROC curve. At 100% specificity (no GH deficient subjects), it had the highest sensitivity, 63.6% (49.2-78.0% CI) at a cut-off of 2.3 ng/m2 (63.6% of GH sufficient subjects had uGH levels > 2.3 ng/m2). When this and other cut-offs were applied to the test group, we found consistency between the observed and predicted numbers of GH sufficient and deficient subjects. CONCLUSIONS: We conclude that urinary GH is a useful test for the diagnosis of GH sufficiency as defined by serum criteria and can be used to reduce significantly the number of serum stimulation tests.