Systematic reviews and meta-analysis of accuracy studies of the desmopressin test in Cushing's syndrome

Autor:	Giampietro, Rodrigo Rosa [UNESP]
Přispěvatelé:	Universidade Estadual Paulista (Unesp), Nogueira, Vania dos Santos Nunes [UNESP]
Jazyk:	portugalština
Rok vydání:	2020
Předmět:	Hipófise Endocrinologia Síndrome de Cushing Pituitary ACTH hypersecretion Cushing Syndrome
Zdroj:	Repositório Institucional da UNESP Universidade Estadual Paulista (UNESP) instacron:UNESP
Popis:	Submitted by RODRIGO ROSA GIAMPIETRO (rodrigo_giampietro@hotmail.com) on 2020-09-10T13:34:52Z No. of bitstreams: 1 Tese Rodrigo Rosa Giampietro versao pós entrega.pdf: 1138436 bytes, checksum: 1c17d93e543983d9bb027d63be234778 (MD5) Approved for entry into archive by ROSANGELA APARECIDA LOBO null (rosangelalobo@btu.unesp.br) on 2020-09-11T17:35:04Z (GMT) No. of bitstreams: 1 giampietro_rr_dr_bot.pdf: 1138436 bytes, checksum: 1c17d93e543983d9bb027d63be234778 (MD5) Made available in DSpace on 2020-09-11T17:35:04Z (GMT). No. of bitstreams: 1 giampietro_rr_dr_bot.pdf: 1138436 bytes, checksum: 1c17d93e543983d9bb027d63be234778 (MD5) Previous issue date: 2020-02-27 Este estudo objetivou avaliar a acurácia do teste da desmopressina em diferenciar doença de Cushing (DC) da síndrome do ACTH ectópico (SAE) e da síndrome de pseudo-Cushing (PC), através de duas revisões sistemáticas de estudos da acurácia de teste diagnóstico (DTA) nos quais pacientes com DC e SAE ou com DC e PC foram submetidos ao teste da desmopressina. As bases de dados LILACS, PubMed, EMBASE, e CENTRAL Cochrane foram pesquisadas. Dois revisores, independentemente, avaliaram os estudos em relação à eligibilidade, risco de viés e extraíram os dados. As meta-análises utilizaram os modelos bivariado ou hierarchical summary receiver operating characteristic (HSROC). A qualidade da evidência foi obtida de acordo com os critérios do GRADE (Grading of Recommendations Assessment, Development and Evaluation Working Group). Cinco estudos de DTA incluídos na revisão sistemática e meta-análise de DC e PC totalizaram 203 pacientes; a maioria deles apresentava hipercortisolismo leve. Com baixa certeza na evidência, a sensibilidade combinada para o diagnóstico de DC foi 89 % (95% Intervalo de Confiança (IC), 81%-94%) e a especificidade 93% (95% IC, 87%-97%). A acurácia diagnóstica obtida pela symmetric summary receiver-operator characteristic (curva ROC) foi 0.96 (95% IC, 0.94 - 0.98). A análise univariada por subgrupos em relação a uma elevação do hormônio adrenocorticotrófico > 4 pmol/L associado a um cortisol basal > 331 nmol/L vs. ACTH > 6 pmol/L resultou em sensibilidade e especifidade de 93 % (83-98 %) vs. 85 % (70-94 %) e de 94 % (83-99%) vs. 94 % (84-99 %), respectivamente, sem diferença estatística. Oito estudos DTAs tiveram 345 indivíduos incluídos na revisão sistemática entre DC e SAE; sete estudos (331 participantes) utilizaram uma elevação percentual no ACTH (Δ%) e sete (175 participantes) utilizaram um Δ% cortisol. A sensibilidade sumária para o diagnóstico de DC foi de 0.83 (95% Intervalo de confiança (IC), 0.78-0.87) para o Δ% ACTH e de 0.80 (95% IC, 0.72-0.85) para o Δ% cortisol. Em ambas análises, o intervalo preditivo 95% (IP) não foi próximo ao IC e 17% (15%-28%) dos portadores de DC seriam perdidos no teste da desmopressina. Contudo, para especificidade em relação ao Δ% ACTH e cortisol, as análises demonstraram elevada heterogeneidade entre os estudos através da análise da hierarchical summary ROC curve plot (IP distante do IC) e pelo resultado obtido no Stata (0.62 (IC, 0.45-0.77) e 0.68 (IC, 0.48-0.83), respectivamente. Apenas um estudo avaliou a sensibilidade e especificidade em relação ao incremente do ACTH associado ao do cortisol, com sensibilidade de 0.65 (IC,0.38-0.86) e especificidade de 0.75 (IC, 0.19 – 0.99). Assim, a especificidade do teste da desmopressina em diferenciar DC de SAE foi heterogênea; apesar da incerteza em relação à sensibilidade ter sido inferior, nossos resultados sugerem que cerca de 20% daqueles com DC serão perdidos pelo teste. PROSPERO números de registros: CRD42018085634 (PC) e CRD42017068317 (SAE) The aim of this study was to evaluate the accuracy of the desmopressin test to differentiate Cushing’s disease (CD) from ectopic ACTH syndrome (EAS) and from Pseudo-Cushing´s syndrome (PCS) For this, two systematic reviews of diagnostic test accuracy (DTA) studies in which patients with CD and ectopic ACTH syndrome or with CD and PCS underwent desmopressin test were realized. The LILACS, PubMed, EMBASE, and CENTRAL Cochrane databases were searched. Two reviewers independently assessed the studies for their eligibility, risk of bias, and extracted data. The meta-analysis used bivariate or hierarchical summary receiver operating characteristic (HSROC) models. The quality of the evidence was generated according to the GRADE (Grading of Recommendations Assessment, Development and Evaluation Working Group) criteria. Five included DTAs studies in the systematic review with CD and PCS comprised 203 individuals; most of them had mild hypercortisolism. With low certainty evidence, pooled sensitivity for CD diagnosis was 89 % (95% Confidence Interval (CI), 81%-94%) and the specificity was 93% (95% CI, 87%-97%). The test accuracy obtained by the symmetric summary receiver-operator characteristic (ROC curve) was 0.96 (95% CI, 0.94 - 0.98). The univariate pooled subgroup analysis according to a net (relative) increment in adrenocorticotropic hormone (ACTH) > 4 pmol/L plus a basal serum cortisol > 331 nmol/L vs. ACTH > 6 pmol/L, showed sensitivities and specificities of 93 % (83-98 %) vs. 85 % (70-94 %) and 94 % (83-99%) vs. 94 % (84-99 %), respectively, without statistical differences. We included eight DTAs studies, with 345 individuals in the systematic review of CD vs. EAS; seven studies (331 participants) used an ACTH percent increment (Δ%), and seven (175 participants) used a cortisol Δ%. The pooled sensitivity for CD diagnosis was 0.83 (95% Confidence Interval (CI), 0.78-0.87) for ACTH Δ% and 0.80 (95% CI, 0.72-0.85) for cortisol Δ%. In both analyses, the 95% prediction intervals (PI) were not near the CI and for individuals with CD, 17% (15%-28%) will be missed by the desmopressin test. However, for specificity in regard Δ% ACTH and cortisol, the analyses showed high levels of heterogeneity between studies, which can be seen from both the hierarchical summary ROC curve plot (PI far from CI) and the Stata output (0.62 (CI, 0.45-0.77) and 0.68 (CI, 0.48-0.83), respectively. Only one study evaluated the sensitivity and specificity of a composite rule requiring an ACTH increase and a cortisol increase, whose values were 0.65 (CI,0.38-0.86) and 0.75 (CI, 0.19 – 0.99), respectively. Thus, the specificity of desmopressin test to distinguish CD from EAS has been very heterogeneous. Although the uncertainty in the sensibility is much lower, our results suggest that an average of 20% of CD diagnosis would be lost by this test. PROSPERO registration numbers: CRD42018085634 (PCS) and CRD42017068317 (EAS)
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=od______3056::2cf256d802485c12525eef5a7ad51921 Zobrazit plný text záznamu