Presentation and diagnosis of childhood-onset combined pituitary hormone deficiency: A single center experience from over 30 years (V体育ios版)

"VSports app下载" Patients

The formation of the patient cohort is shown in Figure 1. We searched the patient registry of Children and Adolescents at Helsinki University Hospital (HUH) for patients with diagnoses of hypopituitarism and related disorders from January 1985 to September 2018 (n=2718, Figure 1). From the search, we identified (i) hits for CPHD ICD-codes (n=237), (ii) familial cases (n=106), and (iii) hits for other hypopituitarism ICD-codes between January 2009 and September 2018 (from the launch of electronic patient records; n=644). Additionally, (iv) we went through growth database search results for HUH catchment area short patients (height ≤-3 SDS at the age of ≥3 years) between 1990 and 2015 (n=785).³⁵ Hits (ii)-(iv) were inspected manually, to identify patients with CPHD but lacking the appropriate diagnosis. Patients with isolated hormone deficiencies, including IGHD, were excluded, to establish a cohort of well-defined CPHD patients.

The identified CPHD cases (n=196+23, Figure 1) were inspected to fulfill the inclusion criteria of at least two biochemically verified pituitary hormone deficiencies (for details, see Supplementary Table 2). Patient records and growth charts of patients with insufficient biochemical data were evaluated by experienced clinicians (PJM, MH, TR). The final study cohort comprised 124 patients (69 boys and 55 girls; Figure 1). Unequivocally subnormal hormone levels related to GH, TSH, and ACTH deficiency were present in 105, 78, and 29 patients, respectively. In others, hormone replacement therapy had been started either (i) based on clinical evaluation (e.g., after pituitary-hypophyseal region tumor resection), (ii) after partial laboratory testing of the pituitary function, or (iii) due to borderline low hormone level, and/or imminent progression of the deficiency in the context of high risk for hypopituitarism. Especially, the inclusion of gonadotropin deficiency and diabetes insipidus often required evaluation by the clinicians. In further analyses of the included patients, hormone replacement therapy is reported as a marker of pituitary hormone deficiency.

Clinical data

All diagnoses of the eligible patients were obtained from the patient records to classify acquired and congenital forms of CPHD. Acquired CPHD was defined to include CPHD of causes extrinsic to pituitary, and congenital CPHD included all idiopathic, syndromic, and genetic CPHD, regardless of the age at presentation. Presenting symptoms for all patients were collected. Age at presentation was the age at first visit or consultation by a HUH pediatric endocrinologist or pediatrician due to findings indicating the need for pituitary assessment. Age at diagnosis was the age at which the physician had recorded the deficiency of at least two pituitary hormones into the patient records. If not available, the date of commencement of hormonal treatment after the diagnosis of two pituitary hormone deficiencies was used.

Additional data collected for the patients with congenital CPHD included (i) birth data (pregnancy details, asphyxia and breech position at delivery, neonatal complications, and features suggesting hypopituitarism),³⁶ (ii) phenotypic features, (iii) findings in brain MRI, (iv) family history of growth disorders and pituitary hormone deficiencies, and (v) auxological data (see next section). Septo-optic dysplasia (SOD) was defined as optic nerve hypoplasia (clinically or radiologically determined) in combination with hypopituitarism (self-evidently present in all our patients) and/or midline brain abnormalities.³⁷ Regarding the neonatal phenotype of hypopituitarism, we categorized hypoglycaemia as a defined feature of CPHD, if the patient had had a hypoglycaemic seizure or apnea, or if she/he had been treated with intravenous glucose for >3 days, or at >3 days of age (persistent hypoglycaemia); other neonatal hypoglycaemia requiring any iv-glucose was defined as a non-specific cue. Similarly, we recorded jaundice as a defined feature, if phototherapy was given at >7 days of age, or with remark “prolonged jaundice”; otherwise, jaundice requiring any phototherapy was defined as a non-specific cue.^36,38 We decided to include the slight, non-specific hypoglycaemia/jaundice phenotypes, since the duration or severity of the symptom had not been recorded for all patients, and, for some patients, the hormone replacement therapy was commenced so early that the persistence of the neonatal phenotype could not be assessed. However, we report the defined phenotypes separately. As for the neonatal genital phenotype, bilateral cryptorchidism and/or micropenis were accepted as signs of gonadotropin deficiency if stated in the patient records (information on penile length was not available for all patients). Preterm infants (born GW<37) were excluded from the neonatal phenotype analyses, since the probability of jaundice (both the severity and duration) and hypoglycemia among them are considerably higher than in full terms.^39,40

We limited the incidence estimate of congenital CPHD to patients born between 2000 and 2018 in the HUH catchment area. The inclusion of patients from earlier years might have been more uncertain in terms of completeness of the patient search results and due to geographical changes in the catchment area. The number of these patients was compared to the number of live-born children in the Helsinki University Hospital catchment area between 2000 and 2018.⁴¹

Growth

Measurements of the patient's height, weight, and bone age were acquired from the electronic growth database (Pediator©; Tilator Oy). The auxological measurements had been carried out by trained nurses. Preterm infants (born GW<37) were excluded from the growth analyses, since catch-up growth in these patients is expected up to the age of 2,3 years, and reduced growth during the first years of life does not per se implicate the need for investigations of underlying pathology.^42,43 To evaluate the growth deflection in patients with CPHD, we used the Finnish growth monitoring rules based on both static (deviation of height SDS (HSDS) from the patient's target height or the general population) and dynamic (deviation of HSDS from the patient's previous measurements) auxological criteria. These criteria are adjusted to detect the 0·5% most slowly and rapidly growing children in the general population for referral to further investigations.⁴⁴ Isolated violation of the growth screening rules was, however, omitted, if subsequent growth was normal.

Genetics

Patients diagnosed with congenital CPHD lacking previous molecular genetic diagnoses were recruited to genetic studies. In total, 21 patients (16 directly recruited from the HUH registry, three based on Helsinki Biobank DNA samples in the HUH registry, and two from Kuopio University Hospital; only the 19 HUH patients were included in the clinical section of this paper), and 21 family members were enrolled. DNA samples from both parents were available for seven indexes, and from one parent for four indexes, while no family genetic data were available for ten indexes (Table 5).

Whole exome sequencing was performed at the sequencing laboratory of the Institute for Molecular Medicine Finland (FIMM) Technology Centre, University of Helsinki. The DNA of the study subjects was extracted from peripheral blood leukocytes and the whole exome sequencing was performed with Illumina Novaseq SP PE150, Novaseq S1 PE150, or Novaseq S1 PE101 technology. First, the adapter was trimmed from the reads, as well as any low-quality nucleotides from the 5’ or 3’ ends of the read, removing pairs with less than 36 bp. The reads were aligned to the GRCh38 reference genome with the BWA (Burrows-Wheeler Aligner). Non-unique read pairs and non-unique single reads were removed and GATK Base Recalibrator was used to clean the alignment. Any potential PCR duplicates were removed using Picard MarkDuplicates, and GATK IndelRealigner was used for indel sites. The mpileup from the SAMTOOLS package was used for variant calling.⁴⁵ The sequencing yielded 90% 20x coverage.

The subjects’ VCP files were annotated using ANNOVAR.⁴⁶ For a variant to be potentially causative and thus eligible for further analysis, we determined that the variant (i) should be in the previously published genes implicated in CPHD/GHD (n=80, Supplementary Table 1), (ii) be either exonic and nonsynonymous, or lie in the consensus splice site or its proximity, which we defined as within ten bases from the splice site, (iii) have a reported minor allele frequency (MAF) equal to or below 2% in all subpopulations in the gnomAD (provided in the ANNOVAR annotations), dbSNP (https://www.ncbi.nlm.nih.gov/snp/), and SISu (http://www.sisuproject.fi/) databases (manually verified).^47,48 Variants fulfilling these criteria were classified according to the ACMG/AMP 2015 guidelines using the InterVar bioinformatics software tool.^49,50

Ethics

The patient studies (direct patient recruitment and the Biobank study) had been approved by the Ethics Committee of Helsinki University Hospital. Genetic studies were carried out according to the Declaration of Helsinki. All directly recruited patients and their guardians gave their written informed consents to participate in the study.

Statistical analyses

The data were analyzed using Microsoft Excel 2016, SPSS statistical software, (SPSS, Version 25.0. Armonk, NY: IBM Corp), and R version 4.0.3. To assess the variable distribution, we visualized the distribution and assessed skewness and kurtosis parameters. Additionally, normality was tested with Shapiro-Wilk test of normality, acknowledging that for small samples, the statistical power of the test may not be sufficient to reject the null hypothesis. Mann-Whitney U-test was used to evaluate the statistical differences in the ages at presentation and diagnosis between the patients with acquired and congenital disease. For auxological data, Mann-Whitney U-test was used to compare Groups I and II for target height, growth screening measures, delay to GH treatment, and latest height measures. T-test was used for birth measures. Wilson score interval was used to estimate the population-level incidence of congenital CPHD. Weighting was not considered necessary for the estimate, due to the representativeness of the patient cohort in both national and international context, and the lack of previous incidence estimates in a similar patient population (for further details, please see Discussion). The data are reported with mean±SD or median interquartile range (IQR) based on the variable distribution.

Role of the funding source

The funders had no role in study design, data collection, data analysis, data interpretation, or writing of the report. All authors had access to study data and approved the manuscript for publication.

Overview of the etiology of CPHD

We investigated the etiologies and diagnostic features of CPHD in the largest tertiary center in Finland. The etiologic spectrum is presented in Table 1. The majority (61%;76/124) of the cases had acquired CPHD, with brain tumors (mainly craniopharyngiomas and gliomas) as the most common etiology. Rare acquired etiologies included iatrogenic, brain trauma, and inflammatory/infiltrative diseases. Also, as we aimed to differentiate developmental mechanisms of hypopituitarism from external causes, one patient with empty sella and hypopituitarism conceivably due to cerebral fluid compression of the pituitary, was classified as acquired CPHD. Congenital CPHD was found in 48 (39%) patients, of whom seven (15%) had a defined genetic diagnosis. One patient with congenital CPHD had died during early childhood despite adequate hormone substitution treatment.²³

According to our inclusion criteria, a retrospectively verifiable biochemical or clinical GH, TSH, or ACTH deficiency, or diabetes insipidus (DI) was present in 120, 109, 69, and 23 of the 124 patients, respectively. However, 60 patients in total received desmopressin treatment for DI since this had often been started based on the treating clinician's interpretation and partial laboratory findings suggesting DI. Sex steroid replacement therapy had been commenced in 52% of patients (n=64, of whom 34 females) at the mean age of 14·1±1·9 years.

The age at presentation and diagnosis for acquired and congenital CPHD are shown in Figure 2. Patients with congenital CPHD presented at a younger age (median 2·4, IQR 0·5 to 5·4 years) compared to patients with acquired CPHD (median 8·9, IQR 5·6 to 12·1 years) p<0·0001. Similarly, the age at diagnosis was lower in patients with congenital (median 3·9, IQR 0·2 to 9·5 years) than acquired (median 9·9, IQR 6·2 to 14·3 years) CPHD p<0·0001.

Phenotypic features of acquired CPHD

Patients with acquired CPHD often presented to medical investigations due to neurologic symptoms of the underlying disease, evident in 54 (71%) patients. These included ophthalmic manifestations (e.g., poor vision), symptoms related to increased intracranial pressure (nausea and headache), deterioration in motor skills, and non-specific symptoms, such as tremor, fatigue, and memory problems. Symptoms related to hypopituitarism (reduced growth, delayed puberty, and/or signs of DI) were among the presenting symptoms in 23 (30%) patients, of whom seven also had neurologic or non-specific symptoms. Of the 67 patients with tumor etiology, CPHD was diagnosed in 19 at presentation and in 32 after brain surgery (often shortly after the first presentation due to tumor-related symptoms but in a few patients after reoperation). In 15 tumor patients, CPHD occurred during follow-up after tumor treatment (including chemotherapy/irradiation). For one patient, the data was not available.

Phenotypic features of congenital CPHD

The deficiency of all anterior pituitary hormones (GH, ACTH, TSH, and FSH/LH) was present in 14 (30%) patients with congenital CPHD (Supplementary Table 3). Other frequent combinations of hormone deficiencies included GH and TSH (n=9, 19%), GH, TSH, and FSH/LH (n=8, 17%), and GH, TSH, and ACTH (n=6, 13%). DI was present in seven (15%) patients. IGHD had preceded the development of CPHD in six (14%) patients (data available in 43/48).

Of the 48 patients with congenital CPHD, eight (17%) were born in breech position, and four (8%) had suffered from birth asphyxia. Neonatal signs associated with hypopituitarism were present in 26 (74%) patients (Figure 3). Six patients were excluded from the neonatal phenotype analysis due to prematurity and seven due to lack of data. A defined neonatal phenotype (see Materials and Methods) was present in 23 (66%) patients. Of these patients, 68% had received a CPHD diagnosis during the first six months of their life (data available in 22/23), including all eight male patients with a genital phenotype. Five patients with defined neonatal signs of hypopituitarism (24%; data available in 21/23), who had either isolated hypoglycaemia or isolated jaundice, did not present to endocrine investigations during the first year of their life.

Of the eight patients with neonatal micropenis/cryptorchidism, six had biochemical evidence of absent minipuberty (tested between the age of 2 and 14 weeks), seven received treatment during minipuberty (five patients with recombinant FSH+testosterone⁵¹ and two with testosterone only). Two patients had been followed up until the time of puberty and required testosterone supplementation, and four patients were still of prepubertal age. All these patients also had GHD as part of their CPHD.

Brain MRI was investigated in 41 (85%) patients with congenital CPHD (Table 2). All of them had abnormalities in the brain MRI. Classic pituitary abnormalities associated with CPHD (anterior pituitary hypoplasia, absent or ectopic posterior pituitary, infundibulum hypoplasia, or the combination of these (pituitary stalk interruption syndrome, PSIS)) were identified in 37 (90%) patients. Extra-pituitary brain anomalies were present in 25 (61%) patients.

Of the patients with neonatal features of CPHD, 22 (85%) had brain MRI investigated. PSIS was present in 14 (64%) and various other combinations of anterior pituitary hypoplasia, infundibulum hypoplasia, and absent/ectopic posterior pituitary in seven (32%) of them, including isolated ectopic posterior pituitary in two. Extra-pituitary MRI findings, present in 15 (68%) included SOD features in 11. Optic nerve hypoplasia was bilateral in 2/5 cases. Four patients (of whom three with SOD) had schizencephaly, local polymicrogyria, local cortical structural abnormality, and/or heterotopic nodules that have previously been associated with, e.g., SOD.⁵² Olfactory bulb hypoplasia was present in three.

Extrapituitary phenotypes were found in 35 (73%) congenital CPHD patients: craniofacial (n=13, 27%), eye (n=16, 33%), genital (n=10, 21%), and developmental delay/epilepsy (n=18, 38%) phenotypes as the most frequent findings (Table 3). Features consistent with SOD were detected in 13 patients with CPHD (Table 1). Their age at CPHD diagnosis was highly variable from early infancy (n=6) to 11-12 years of age (n=3). Eight out of the 12 SOD patients with available patient data had presented with neonatal features of hypopituitarism, and six of them had received a neonatal CPHD diagnosis.

Growth of patients with congenital CPHD

Height growth was analyzed in 26 (54%) patients with congenital CPHD; other patients were excluded due to prematurity (n=6), missing data (n=10), and start of GH Rx treatment without preceding abnormal height growth according to the Finnish growth screening rules (n=6). Height growth deflection was evident in 12 (46%) patients in early infancy (0–6 months of age; Group I) and in 11 (42%) patients between six months to four years of age (Group II). Key auxological measurements for Group I and II are presented in Table 4.

Birth HSDS, birth weight for height, and target height of patients with congenital CPHD were close to the general population (Table 4). Defined neonatal features of hypopituitarism were present in 8/12 patients of Group I and in 5/11 patients of Group II. Further, these were among the presenting symptoms to endocrine investigations in 7/7 patients of Group I and 1/4 patients of Group II, for whom the data were available (data not shown). This suggests that the patients with earlier growth retardation (Group I) had a more severe CPHD phenotype in early infancy. Also, none of the patients in Group II had presented with a neonatal genital phenotype. Reduced height growth was the presenting symptom in endocrine investigations of 11/23 (42%) congenital CPHD patients; 3/12 (25%) in Group I and 8/11 (73%; the only presenting symptom in 45%) in Group II.

The median delay between the first violation of the growth screening rules and the initiation of GH Rx treatment among all congenital CPHD patients was 2·2 years (IQR 1·2 to 3·7 years; n=19). There was no statistical difference in the median delay to GH Rx between Group I (1·9, IQR 0·6 to 6·1 years) and Group II (2·5, IQR 1·2 to 3·2 years), p=0·778. Figure 4 presents the height growth of patients in Group I and Group II. To more reliably estimate the GHD diagnostic delay, we analyzed the delay between the violation of the growth screening rules and the presentation at pediatric endocrinologist for pituitary assessment, in the subgroup of patients referred to investigations due to divergent growth. This delay was median 2·2 years (IQR 1·2 to 3·6 years, n=11). Only two of these patients presented for evaluation of the pituitary function during the year following the growth screening alert.

Incidence estimate for congenital CPHD

During the years 2000-2018, 422 258 live children were born in the Helsinki University Hospital catchment area⁴¹ and 26 patients born in this region during the same period were diagnosed with congenital CPHD, giving an estimated incidence of 26/422 258 (1 in 16 000 live-born children, 95%CI 1/11 000-1/24 000) for congenital CPHD.

Genetics

In previous clinical or research investigations, a molecular genetic cause of CPHD had been identified in seven patients: a homozygous deletion of PROP1 in two siblings of Kurdish origin with CPHD, compound heterozygous loss-of-function variants in TBC1D32 in siblings with PSIS, and mild craniofacial dysmorphism; a previously described loss-of-function variant in SOX2 in a female patient with SOD (anophthalmia and CPHD), a Xq27.1 microduplication of unknown size leading to loss of SOX3 in a male patient with PSIS, and a frameshift variant in OTX2 in a male patient with PSIS and retinitis pigmentosa.^23,53, 54, 55

Of the 21 patients who were enrolled in genetic investigations, 18 carried rare (MAF≤2%) nonsynonymous or consensus splice site variants in genes implicated in CPHD/GHD, as shown in Table 5. Due to incomplete sequence data availability from family members, we were unable to analyze the mode of inheritance for most variants. Variants with MAF either ≤0·5% or not reported, were identified in ten indexes in 16 different genes (ALMS1, ARNT2, CDON, CHD7, CLCNKB, DCHS1, FGFR1, IGSF10, ISL1, KAT6A, L1CAM, LAMB2, LHX3, MAGEL2, RBM28, and SHH). However, according to the ACMG/AMP 2015 guidelines, classifications for most variants were “Uncertain significance” or “likely benign”. The rare variants identified in our study were all heterozygous, and, considering the incidence estimate of CPHD presented above, the possibility of an autosomal dominant monogenic disease could only be considered in patients carrying extremely rare variants. Limiting to variants with MAF≤0·01% or not reported and excluding those classified as “benign” or “likely benign”, we encountered variants in ISL1 (patient #10); SHH (patient #12); ALMS1 and L1CAM (patient #6); and FGFR1, CHD7, and SHH (patient #7). The only variant classified as “likely pathogenic'', was the heterozygous SHH c.676G>A, p.(Ala226Thr) variant carried by patient #12 with PSIS (GH, ACTH, TSH deficiencies, and estrogen treatment due to suspected partial gonadotropin deficiency) and her unaffected mother.