Adult height in growth hormone (GH)-deficient children treated with biosynthetic GH. The Genentech Growth Study Group

Near-adult height (AH) was determined in 121 children (72 males and 49 females) with GH deficiency (GHD) who were prepubertal when they began treatment with recombinant DNA-derived preparations of human GH. AH as a SD score was -0.7 +/- 1.2 (mean +/- SD), significantly greater than the pretreatment height SD score (-3.1 +/- 1.2), the predicted AH SD score (-2.2 +/- 1.2; Bayley-Pinneau method), and the height SD score at the start of puberty (-1.9 +/- 1.3). In contrast to studies of GH treatment outcome, which used pituitary-derived GH (pit-GH) in lower doses, we found that males did not have a higher AH SD score than females, spontaneous puberty did not diminish AH, and AH was significantly greater than that predicted at the start of GH treatment. In a multiple regression equation, the statistically significant variables (all P < 0.0001) related to AH (r2 = 0.70) were the following: duration of treatment with GH, sex (males were taller than females, as expected for the normal population), age (younger children had a greater AH) and height at the start of GH, and growth rate during first year of GH. For the AH SD score (r2 = 0.47), pretreatment predicted AH, duration of GH, and bone age delay were significant (P < 0.0002) explanatory variables. Bone age delay (chronological age-bone age) had a negative impact on the AH SD score. Target height, etiology of GHD, previous treatment with pituitary GH, and the presence or absence of spontaneous puberty did not significantly improve the prediction of AH. Early diagnosis of GHD and continuous treatment with larger doses of GH to near AH should improve the outcome in children with short stature due to GHD.     

Moral assessment of growth hormone therapy for children with idiopathic short stature

Subjects ranging in age from 50 to 89 years old, either with or without dementia were studied by both ELISA for anti-human T-cell lymphotropic virus type I (HTLV-I) gag 100-130 antibody and by cranial CT in order to clarify the relationship between HTLV-I infection and dementia. The frequency of anti-HTLV-I antibody was found to be significantly higher in the patients with dementia (24/130, 18.5%) than in those without dementia (11/139, 7.9%) (P=0.0169). Among the various types of dementia, HTLV-I seropositivity was found to be significantly associated with vascular dementia (11/48, 23%) (P=0.0087), but not with Alzheimer type dementia. In addition, HTLV-I seropositivity was also associated with Babinski sign, and the severity of cerebral infarction, ventricular dilatation and periventricular lucency on CT. The presence of HTLV-I therefore appears to be one of the risk factors for vascular dementia in HTLV-I endemic areas.     

Short- and long-term (final height) data in children with normal variant short stature treated with growth hormone

Seventeen children with normal variant short stature and a predicted height below -2 SDS were treated with growth hormone (GH) six times a week for a period of 5 years. Patients were randomly selected to receive three different doses of GH, group 1 (n = 6) 3 IU/m2 per day, group 2 (n = 6) 4.5 IU/m2 per day and group 3 (n=5) 3 IU/m2 per day in the 1st year and 4.5 IU/m2 per day thereafter. There was a significant increase in height after 1 and 2 years for all patients and for all subgroups. However, this increase was not dependent on GH dose. The decrease in height velocity during the 2nd year was not prevented by the increase of GH dose in group 3. The change of predicted height after 2 years was +0.75 SDS (according to Tanner Whitehouse). Fourteen children have been treated for 4 years and 8 children for 5 years without a further change in height prediction. Nine patients have reached final height which was 2.4 cm (+0.41 SDS) above pretreatment height prediction. Final height was nearly identical to predicted height after 1 year of therapy. CONCLUSION: An increment in height prediction was observed during the first 2 years of GH treatment and maintained thereafter. However, there was only a minor increase in final height over predicted height which does not justify the general use of GH in children with normal variant short stature.     

Isolated partial growth hormone deficient short stature with syringomyelia not associated with birth injury

A 59-year-old female with 20-year history of slowly progressing muscle atrophy and sensory disturbance of upper extremities showed short stature, scoliosis, hunger type of sensory dissociation of the upper extremities and pyramidal tract sign of the lower extremities. Magnetic resonance imaging (MRI) clarified hypoplasia of the anterior pituitary lobe, Arnold-Chiari malformation and cervical syringomyelia. Insulin and arginine stimulating tests revealed partial type of isolated growth hormone (GH) deficiency but GH gene analysis detected no defects of GH genes. It was considered to be a rare case of non-hereditary hypopituitarism with Chiari malformation and syringomyelia not associated with perinatal injury, namely a midline anomaly syndrome.     

Recombinant human erythropoietin (rhEPO) treatment potentiates growth hormone (GH) response to growth hormone releasing hormone (GHRH) stimulation in hemodialysis patients

On the basis of previously described effects of recombinant human erythropoietin (rhEPO) treatment on endocrine abnormalities present in uremia, we assessed the possible effect of treatment with rhEPO on growth hormone (GH) response to growth hormone releasing hormone (GHRH) in a group of uremic patients. Eight patients on maintenance hemodialysis for 12 to 228 months, not previously treated with rhEPO, were tested with 100 micrograms of GHRH i.v. in bolus before and after three months of rhEPO treatment (40 U/kg i.v. three times a week). Before treatment, the GH response to GHRH was characterized, in uremic patients, by remarkable differences in plasma GH values and in the pattern of response curve in single patients. The variability of GH response was not modified after rhEPO treatment; however, an overall potentiation of GH response with a significant increase of plasma GH (p = 0.017 at 15 min, p = 0.035 at 30 min after GHRH injection) was observed in the tests performed after treatment. rhEPO administration induced an evident improvement of anemia, blood hemoglobin concentration being 5.3-7.6 g/dl before and 9.1-11.3 g/dl after treatment; however a demonstrable correlation between the potentiation of GH response to GHRH and the increase of hemoglobin concentration was not observed.     

Dental maturity in children of short stature, with or without growth hormone deficiency

Macrophages are important constituents of the immune system by exerting phagocytosis on invading pathogens as well as secreting various immunoregulatory factors. Generation of human macrophage hybridoma has not been possible so far due to the lack of an appropriate fusion partner cell line. In the present study, an 8'-azaguanine resistant cell line, termed HL-60R, was established by drug selection of the promyelocytic cell line HL-60. This novel cell line showed resistance to high concentrations of 8'-azaguanine and was sensitive to aminopterin. These characteristics make it suitable for serving as a potential fusion partner cell line in the development of macrophage hybridoma. Cell-surface analysis by FACS revealed that HL-60R cells per se do not express MHC-class II molecules or the macrophage marker, CD11b. PEG-mediated fusion of HL-60R was performed with PBMC-derived human macrophages. Fluorescence labelling of ex vivo isolated macrophages prior to fusion and subsequent FACS analysis showed that PEG-4000 is a more effective fusion agent than PEG-1500. The generation of this novel fusion partner cell line opens the possibility for development of human macrophage hybridoma or other cell lines from myelocytic origin. Such hybridoma clones will not only enable a more convenient study of these cell but will also provide an excellent host site for the proper production and expression of various recombinant proteins from myelocytic origin in vitro.     

GH and GnRH analog treatment in children who enter puberty at short stature

It is well known that height at the onset of puberty is closely related to final height. To improve final height of short children who enter puberty at short stature, twenty-one short boys and six short girls were treated with a combination of GH and GnRH analog. The boys started the combination treatment at a mean age of 12.0 years when their mean height was 128.5 cm (-2.74 SD) and the girls at a mean age of 10.68 years when their mean height was 126.4 cm (-2.23 SD). The boys discontinued GnRH at a mean age of 16.88 years after a mean treatment period of 4.89 years when their height was 153.7 cm (-2.75 SD), and the girls at a mean age of 13.89 years after a mean treatment period of 3.20 years when their height was 143.3 cm (-1.94 SD). Bone age maturation significantly decelerated during the combination treatment. Bone age rarely exceeded 14 years in boys and did not exceed 13 years in girls. Bone age maturation during combination treatment decelerated after bone age 12 years in boys and 10.5 years in girls. On average, bone age matured at a mean rate of 0.48 years a year in boys and 0.56 years a year in girls during the combination treatment. During the combination treatment, height velocity did not decelerate rapidly and remained at 3-5 cm/year for a longer duration because of the bone age deceleration, although a definite pubertal growth spurt was not observed. As a consequence, the mean projected height SDS for bone age increased 1.50 (+/- 0.76) SD in boys and 1.24 (+/- 0.49) SD during the combination treatment. Although most of the patients have not yet reached their final height, combined GnRH analog and GH treatment should increase the pubertal height gain and the adult height in short children who enter puberty early for height, when the post-GST growth is taken into account. The combination treatment seems more effective in boys than in girls. This improvement is attributed to the lengthening of the treatment period by slower bone maturation and maintained growth velocity.     

Effects of short- and long-term dexamethasone treatment on growth and growth hormone (GH)-releasing hormone (GRH)-GH-insulin-like growth factor-I axis in conscious rats

Although the inhibitory effects of a chronic excess of glucocorticoids (GC) on body growth and GH secretion are well established, the mechanisms involved remain unclear. In this study, we examined the chronic effects of a high dose of dexamethasone (DEX) on spontaneous GH secretion and insulin-like growth factor (IGF)-I in conscious rats. The animals were given daily i.p. injections of DEX (200 microg/day) for either one or four weeks. Body growth assessed by tibia length and serum IGF-I levels was significantly inhibited 1 week after treatment. By contrast, spontaneous GH secretion was not altered 1 week after the treatment. Neither hypothalamic GRH and somtatostain mRNA levels nor GH responses to GRH from single somatotropes were affected 1 week after the treatment. Four weeks after DEX treatment, body growth of the rats was noticeably suppressed. Interestingly, spontaneous GH secretion, hypothalamic GRH mRNA levels and GH responses to GRH were all inhibited 4 weeks after treatment. Pituitary GRH receptor mRNA levels were not altered 1 week after treatment, but increased after 4 weeks. These results indicate that a high dose of DEX initially impairs IGF-I production and subsequently inhibits spontaneous GH secretion in rats. Inhibition of spontaneous GH secretion resulting from chronic GC excess is due, at least in part, to the impairment of hypothalamic GRH synthesis and pituitary GH responsiveness. An increase in the pituitary GRH receptor may be caused by decreased GRH secretion.     

Arm span as measurement of response to growth hormone (GH) treatment in a group of children with meningomyelocele and GH deficiency

Children with meningomyelocele (MMC) frequently have impaired linear growth. A number have associated structural brain defects with resultant GH deficiency (GHD). Reproducible measurements of height or length in MMC patients are often hampered by lower limb contractures, spasticity, and scoliosis. Arm span has been proposed as a more reproducible measure of linear growth. Five MMC children documented to have GHD were treated with recombinant human GH (hGH) for 1-3 yr. Their height, arm span, and growth velocity were compared with 32 children with idiopathic GHD treated similarly with hGH. These measures are compared with normal children by being expressed as standard deviation scores. The results of this study indicate that arm span measurements in GHD MMC patients are almost identical to height measurements in idiopathic GHD patients both before and during hGH therapy. The physical condition of children with MMC makes reproducible longitudinal height measurements difficult. Routine determinations of arm span measurements for children with MMC will assist in recognizing growth failure as well as monitoring treatment results.     

Pretreatment with somatostatin analog SMS 201-995 potentiates growth hormone (GH) responsiveness to GH-releasing factor in short children

Previous studies in children have shown inconsistent, poorly reproducible GH responses to exogenous GH-releasing factor (GRF), with wide individual variability. In the present study, we tested the hypothesis that prior administration of the long-acting somatostatin analog, SMS 201-995 (SMS), will enhance GH responsiveness to a subsequent GRF challenge. Two study protocols were employed in 37 children with short stature [M = 31, F = 6, ages 11.8 +/- 1.6 yr (mean +/- SEM), height -2.25 +/- 0.55 SDS (SD scores)]. In both studies, each subject served as his/her own control. In the first study, which was designed to determine optimal SMS dose and regimen, SMS, in doses ranging from 0.8-2.2 micrograms/kg sc, was randomly administered or omitted at 0800 h after an overnight fast, and a GRF bolus (50 micrograms, iv) was given 4 h later. In the second study, we employed a protocol identical to study 1 except for the use of standard doses of SMS (1 microgram/kg, sc) and GRF (1 microgram/kg, iv) and an additional 1-h delay of the GRF injection. Plasma GH levels were measured every 20 min from 0800 h until 2 h after the GRF injection in both studies. In study 1 (n = 12; M = 10, F = 2), SMS significantly suppressed spontaneous GH secretion (expressed as the mean +/- SEM GH AUC during the 4-h SMS-GRF interval, AUC 1:2.2 +/- 0.4 vs. 6.2 +/- 0.9 micrograms/L.h; P < 0.001), GH responsiveness to GRF (GH AUC during the 2 h after the GRF injection, AUC 2: 41.5 +/- 7.8 vs. 85.0 +/- 13.5 micrograms/L.h; P < 0.001), and the GH peak response (17.4 +/- 3.1 vs. 36.0 +/- 6.2 micrograms/L; P < 0.001), compared to control tests. In contrast, in study 2 (n = 25; M = 21, F = 4), whereas spontaneous GH secretion was still suppressed during the 5-h SMS-GRF interval (AUC 1:3.8 +/- 0.4 vs. 7.4 +/- 1.1 micrograms/L.h; P < 0.001), both the GH peak response (56.7 +/- 5.5 vs. 30.5 +/- 3.0 micrograms/L; P < 0.0001) and the GH AUC (AUC 2: 103.7 +/- 10.3 vs. 77.5 +/- 6.8 micrograms/L.h; P < 0.05) after GRF administration were significantly augmented by pretreatment with SMS, compared to control tests. Taken together, these results indicate that a priming SMS dose of 1 microgram/kg has a significant permissive effect on GH responsiveness to exogenous GRF administered 5 h later.(ABSTRACT TRUNCATED AT 400 WORDS)     

Skeletal effects of two years of treatment with low physiological doses of recombinant human growth hormone (GH) in patients with adult-onset GH deficiency

A low bone mass in adults with childhood-onset GH deficiency (GHD) is likely to be caused by deficient bone accretion during childhood and early adulthood, whereas a decreased bone mass in patients with adult-onset GHD is likely to be caused by an imbalance in bone remodeling. Data on bone mineral density (BMD) and biochemical parameters of bone metabolism and data on response of these parameters to treatment with GH are scarce in patients with adult-onset GHD. It has been suggested that in patients with GHD, GH at the relatively high dose originally used may have beneficial effects on the skeleton. To address the question as whether lower, more physiological doses would have similar effects on the skeleton, we studied 47 patients with adult-onset GHD (27 women and 20 men, range 26-70 yr) randomized to receive one of three recombinant human GH (rhGH) dose regimens: 0.6 IU/day, 1.2 IU/day, or 1.8 IU/day as part of a study examining optimal GH dose replacement therapy. After 24 weeks of treatment, the dose of rhGH was individually adjusted to maintain the concentration of serum insulin growth factor-I within the normal laboratory reference range. Biochemical parameters of bone metabolism were measured at baseline and after 24 and 52 weeks and 2 yr of treatment. BMD of the lumbar spine was measured at baseline and after 52 weeks and 2 yr of treatment. Parameters of bone metabolism generally fell within the low-normal range and increased in a dose-dependent manner at 24 weeks of treatment. Between 24 and 52 weeks of rhGH treatment, mean serum osteocalcin levels and alkaline phosphatase activity further increased, whereas mean 24-h urine hydroxyproline/creatinine and N-telopeptide/creatinine excretion remained unchanged. After 52 weeks of treatment, serum alkaline phosphatase activity and 24-h urine hydroxyproline/ creatinine excretion decreased, although not to pretreatment levels. Mean BMD at the lumbar spine (Z-score) was normal at baseline (-0.20 +/- 0.16) and increased during treatment (at 2 yr of treatment: 0 +/- 0.20; P < 0.005). Our data suggest that a low physiological dose of rhGH, individually adjusted to maintain serum insulin-like growth factor I levels within the normal laboratory reference range, increased bone turnover in favor of bone formation, as suggested by the significant, albeit small increase in BMD observed after 2 yr of treatment. Further studies are required to establish whether in patients with adult-onset GHD the preservation and/or increase in bone mass observed with the use of physiological doses of rhGH could be maintained with longer-term treatment.     

Effects of luteinizing hormone-releasing hormone analog-induced pubertal delay in growth hormone (GH)-deficient children treated with GH: preliminary results

To study the effect of delaying epiphyseal fusion on the growth of GH-deficient children, we studied 14 pubertal, treatment naive, GH-deficient patients (6 girls and 8 boys) in a prospective, randomized, placebo-controlled trial. Chronological age was 14.5 +/- 0.5 yr, and bone age was 11.6 +/- 0.3 yr (mean +/- SEM) at the beginning of the study. Patients were assigned randomly to receive GH and LH-releasing hormone (LHRH) analog (n = 8) or GH and placebo (n = 6) during 3 yr, with planned continuation of GH treatment until epiphyseal fusion. Patients were measured with a stadiometer and had serum LHRH tests, serum testosterone (boys), serum estradiol (girls), and bone age performed every 6 months. Patients treated with GH and LHRH analog showed a clear suppression of their pituitary-gonadal axis and a marked delay in bone age progression. We observed a greater gain in height prediction in these patients than in the patients treated with GH and placebo after 3 yr of treatment (mean +/- SEM, 14.0 +/- 1.6 vs. 8.0 +/- 2.4 cm; P < 0.05). These preliminary findings suggest that delaying epiphyseal fusion with LHRH analog in pubertal GH-deficient children treated with GH increases height prediction and may increase final height compared to treatment with GH alone.     

Serum growth hormone (GH) binding protein, IGF-I and IGFBP-3 in patients with beta-thalassaemia major and the effect of GH treatment

OBJECTIVE: Levels of IGFI have been shown to be low in transfusion-dependent thalassaemia and there is preliminary evidence to suggest that this may be reversed by GH treatment. In this further study we have evaluated serum growth hormone (GH) binding protein (GHBP), IGF-I and IGFBP-3 in patients with beta-thalassaemia major and the effects of GH treatment on these various parameters. PATIENTS: Fifty-six transfusion dependent patients with beta-thalassaemia major without GH deficiency between 2 and 20 years of age were studied. Thirteen non-GH deficient patients with heights of -1.5 SD or more were treated with GH at a dose of 0.14 IU/kg/day subcutaneously for 1 year. MEASUREMENTS: Serum GHBP, IGF-I and IGFBP-3 were measured in all the patients. In the 13 patients treated with GH, these serum parameters were measured before and after 3, 6 and 12 months of treatment. RESULTS: The mean serum GHBP concentrations were normal in both prepubertal and pubertal children but the serum IGF-I and IGFBP-3 concentrations were low throughout childhood and adolescence. There was a significant correlation between serum IGF-I and IGFBP-3 concentrations (r = 0.79; P = 0.0001) but there was no correlation between the height SDS of the patients with serum GHBP, IGF-I or IGFBP-3 levels. GH treatment in the 13 patients resulted in significant growth acceleration associated with a significant rise in the serum IGF-I and IGFBP-3 and a significant fall in serum GHBP concentrations. CONCLUSIONS: The low serum concentrations of IGF-I and IGFBP-3 in the presence of normal GH reserve and serum GHBP concentrations in patients with beta-thalassaemia suggest a state of partial GH insensitivity at the post-receptor level. This partial GH insensitivity state can be overcome by supraphysiological doses of exogenous GH. The lack of correlation of IGF-I, IGFBP-3 and GHBP with height SDS of the patients imply that the growth failure commonly observed in patients with beta-thalassaemia major may not be specifically related to dysregulation of the GH-IGF-I axis. GH therapy resulted in significant increase in serum IGF-I and IGFBP-3 but a significant fall in GHBP.     

Co-expression of bovine growth hormone (GH) and human GH antagonist genes in transgenic mice

Bovine growth hormone (bGH) transgenic (Tg) mice have been shown to possess enhanced growth phenotypes and exhibit severe glomerulosclerosis. One amino acid substitution in GH, i.e. G119R in bGH or G120R in human (h) GH, results in GH antagonists (GHAs). GHA-Tg mice exhibit dwarf phenotypes and normal kidneys. In order to investigate the possibility of GHAs as pharmaceutical agents for the treatment of human diseases with excessive GH levels, we cross bred mice that express bGH with those that express hGHA. Double positive Tg mice were identified that express both genes although at different levels. Kidney histological studies revealed that the double positive Tg mice with high GHA/GH expression ratios possessed normal or near normal kidneys, whereas those with low GHA/GH ratios exhibited glomerulosclerosis similar to GH-Tg mice. Thus, co-expression of GH and GHA genes in vivo results in animal phenotypes and kidney histopathologies which are a reflection of the relative expression levels of each gene.     

Novel compound heterozygous mutations of growth hormone (GH) receptor gene in a patient with GH insensitivity syndrome

A girl with severe growth retardation had the clinical features of Laron syndrome. Her serum insulin-like growth factor-I level was completely unresponsive to exogenous GH administration. The serum GH-binding protein (GHBP) level was below the detectable limit in the patient, but it was normal in her parents and brother. Her parents and brother were normal in their height. Amplification with PCR and direct sequencing of her GH receptor gene revealed compound heterozygous mutations. The allele from her mother contained a transversion of G to T in exon 7 that could introduce a stop codon in place of a glutamic acid at amino acid 224. Another mutation was found in the allele in her father and also identified in her brother. It was a C deletion at position 981 in exon 10 that could introduce a frame shift, thereby causing the production of 20 novel amino acids (310-329) instead of the wild-type sequence, the premature termination at codon 330, and the subsequent deletion of the C terminal portion of the intracellular domain. RT-PCR of her father's lymphocytes and sequencing of its complementary DNA revealed that only the wild-type GH receptor messenger RNA was expressed in his lymphocytes, though the mechanism remains unclear. These results suggest that neither of the mutant alleles could generate a functional GH receptor, which would be consistent with the patient's severe growth retardation and undetectable serum GHBP.     

Changes in bone mineral density, body composition, and lipid metabolism during growth hormone (GH) treatment in children with GH deficiency

Adults with childhood onset GH deficiency (GHD) have reduced bone mass, increased fat mass, and disorders of lipid metabolism. The aim of the present study was to evaluate bone mineral density (BMD), bone metabolism, body composition, and lipid metabolism in GHD children before and during 2-3 yr of GH treatment (GHRx). Forty children with GHD, mean age 7.9 yr, participated in the study of bone metabolism and body composition; and an additional group of 17 GHD children, in the study of lipid metabolism. Lumbar spine BMD, total body BMD, and body composition were measured with dual-energy x-ray absorptiometry. Volumetric BMD (bone mineral apparent density, BMAD) was calculated to correct for bone size. BMD, BMAD, lean tissue mass, bone mineral content, fat mass, and percentage body fat were expressed as SD scores (SDS), in comparison with normative data of the same population. Lumbar spine BMD and BMAD and total body BMD were all decreased at baseline. All BMD variables increased significantly during GHRx, lumbar spine BMD SDS, already after 6 months of treatment. Lean tissue mass SDS increased continuously. Bone mineral content SDS started to increase after 6 months GHRx. Fat mass SDS decreased during the first 6 months of GHRx and remained stable thereafter. Biochemical parameters of bone formation and bone resorption did not differ from normal at baseline and increased during the first 6 months of GHRx. Serum 1,25 dihydroxyvitamin D increased continuously during GHRx, whereas PTH and serum calcium remained stable. Lipid profile was normal at baseline: Atherogenic index had decreased and apolipoprotein A1(Apo-A1) had increased after 3 yr of treatment. In conclusion, children with GHD have decreased bone mass. BMD, together with height and lean tissue mass, increased during GHRx. GHRx had a beneficial effect on lipid metabolism.     

Biosynthetic growth hormone in the treatment of low stature in Turner syndrome

BACKGROUND AND STUDY AIMS: The presumptive diagnosis of Candida esophagitis has been included in the Centers for Disease Control (CDC) case definition for full-blown AIDS since 1987. Endoscopic examination should be reserved for patients showing symptoms despite treatment. The purpose of this study was to assess the degree of diagnostic accuracy of the CDC presumptive clinical criteria and to determine the usefulness of upper digestive endoscopy in the diagnosis of Candida esophagitis in patients infected with HIV-1, with and without a previous AIDS-defining event. PATIENTS AND METHODS: A total of 144 HIV-1 infected patients who had undergone an upper digestive endoscopy were studied retrospectively. To determine the risk and the predictive value of the clinical markers, only the 84 patients without prior antimycotic therapy were included. RESULTS: Of the 84 patients without previous treatment, 34 (41%) had a history of an AIDS-defining illness. Candida esophagitis was found on endoscopy in 11 of the AIDS and 28 of the non-AIDS cases. Oral thrush, either alone (relative risk [R.R.] 9.4; 95% C.I. 2.4-36.4; p < 0.01; positive predictive value [PPV] 82%) or in combination with esophageal symptoms (R.R. 7.4; 95% C.I. 2.5-21.9; p < 0.01; PPV 89%), was a reliable marker of Candida esophagitis only in patients with a previous AIDS-defining event. The diagnostic value of the CDC presumptive pattern was confirmed by a multivariate analysis after controlling for the CD4 cell count (R.R. 9.3; 95% C.I. 2.3-25.3; p < 0.01). On the other hand, in HIV-1 positive patients without a previous AIDS-defining event, the diagnostic accuracy of oral candidiasis, either alone (R.R. 1.4; 95% C.I. 0.8-2.4; p n.s.; PPV 64%) or in combination with esophageal symptoms (R.R. 1.1; 95% C.I. 0.7-1.8; p n.s.; PPV 60%), was too low to allow a reliable diagnosis of Candida esophagitis. CONCLUSIONS: A presumptive diagnosis of Candida esophagitis on the basis of the CDC clinical criteria is a valid diagnostic method only in HIV-1 infected patients with a previous diagnosis of full-blown AIDS. Upper digestive endoscopy should be performed in symptomatic patients with no history of an AIDS-defining illness, especially if the diagnosis of esophageal candidiasis is important for surveillance purposes.