Diagnostic accuracy of total serum bile acids or individual bile acids for intrahepatic cholestasis of pregnancy in woman claiming pruritus

Review question
To assess and compare the diagnostic accuracy of total serum bile acids (TSBA) and some components of serum bile acid profile for the diagnosis of intrahepatic cholestasis of pregnancy in woman with onset of pruritus during pregnancy.

Background
'Diagnostic accuracy' means how well a test correctly identifies or rules out disease and informs subsequent decisions about treatment. Intrahepatic cholestasis of pregnancy is a pregnancy-specific liver disorder, in which bile (a digestive fluid) builds up in the liver, impairing the liver (intrahepatic) function. Intrahepatic cholestasis of pregnancy is possibly associated with an increased risk of premature delivery and fetal death, which seems to occur most often during the last weeks of pregnancy. This is why most clinicians choose to induce early delivery of the baby.

In clinical practice, presence of severe pruritus (itchiness) during late pregnancy and 'otherwise unexplained' abnormalities in serum liver tests, seems enough to support the diagnosis of intrahepatic cholestasis of pregnancy. However, excluding all other possible underlying diseases is not always easy; hence confirmation of the intrahepatic cholestasis of pregnancy diagnosis may be possible only after delivery, when spontaneous disappearance of pruritus and improvement of liver tests on blood exams usually occur.

Total serum bile acids (TSBA) are the most used biomarkers for intrahepatic cholestasis of pregnancy in clinical practice. Some components of the serum bile acid profile might provide more specific information than total serum bile acids when diagnosing the disease, defining its severity and monitoring its response to treatment.

Study characteristics
This review considered all evidence provided by studies that assess the diagnostic accuracy of total serum bile acids (TSBA) and any component of serum bile acid profile for intrahepatic cholestasis of pregnancy in woman claiming onset of pruritus during pregnancy.

We assessed all available reports from a wide, systematic search of databases of medical literature, irrespective of design, publication status, language, and study design. We finally included 16 studies, most of them assessing the accuracy (sensitivity and specificity) of TSBA with a cut-off of 10 μmol/L. Most studies had a case-control design, and these studies could have overestimated the diagnostic accuracy.

Key results
When considering the studies with a cut-off of 10 μmol/L for TSBA serum concentration, TSBA overall sensitivity (the ability to correctly identify women with the disease) ranged from 72% to 98% and specificity (the ability to correctly identify women without the disease) ranged from 81% to 97%. However, after performing two different analyses excluding studies with probably less reliable results, the diagnostic accuracy seemed lower. We calculated the overall accuracy also of some components of serum bile acid profile, but the small number of studies and the high variability of the results led to very imprecise data.

Quality of the evidence
Only one of the 16 included studies was performed and reported well (low risk of bias). The remaining 15 studies had problems with study design or reporting (high risk of bias). Only five studies seemed to show low concern regarding applicability of the results in clinical practice.

Conclusions
The overall high risk of bias, the existing concern regarding applicability of the results in clinical practice, and the poor uniformity of our results in the included studies prevents us from making recommendations and reaching definitive conclusions at present. Thus, we do not find any compelling evidence to recommend or refute the routine use of any of these tests in clinical practice. So far, the diagnostic accuracy of TSBA for intrahepatic cholestasis of pregnancy might have been overestimated. There were too few studies to permit a precise estimate of the accuracy of serum bile acid profile components. Further primary clinical research is mandatory. We need both further phase II and phase III diagnostic studies.

Authors' conclusions: 

The overall high risk of bias, the existing concern regarding applicability of the results in clinical practice, and the great heterogeneity of the results in the included studies prevents us from making recommendations and reaching definitive conclusions at the present time. Thus, we do not find any compelling evidence to recommend or refute the routine use of any of these tests in clinical practice. So far, the diagnostic accuracy of TSBA for intrahepatic cholestasis of pregnancy might have been overestimated. There were too few studies to permit a precise estimate of the accuracy of serum bile acid profile components. Further primary clinical research is mandatory. We need both further phase II and phase III diagnostic studies.

Read the full abstract...
Background: 

Intrahepatic cholestasis of pregnancy is a pregnancy-specific liver disorder, possibly associated with an increased risk of severe fetal adverse events. Total serum bile acids (TSBA) concentration, alone or in combination with serum aminotransferases, have been the most often used biomarkers for the diagnosis of intrahepatic cholestasis of pregnancy in clinical practice. Serum bile acid profile, composed of primary or secondary, conjugated or non-conjugated bile acids, may provide more specific disease information.

Objectives: 

To assess and compare, independently or in combination, the diagnostic accuracy of total serum bile acids or serum bile acids profile, or both, for the diagnosis of intrahepatic cholestasis of pregnancy in pregnant women, presenting with pruritus. To define the optimal cut-off values for components of serum bile acid profile; to investigate possible sources of heterogeneity.

Search strategy: 

We searched the Cochrane Hepato-Biliary Group Controlled Trials Register, the Cochrane Hepato-Biliary Group Diagnostic Test Accuracy Studies Register, the Cochrane Library, MEDLINE Ovid, Embase Ovid, Science Citation Index Expanded, Conference Proceedings Citation Index – Science, BIOSIS, CINAHL, two Chinese databases (CKNI, VIP), Latin American and Caribbean Health Sciences Literature (LILACS), Scientific Electronic Library Online (SciELO), Evidence Search: Health and Social Care by the National Institute for Health and Care Excellence (NICE), the World Health Organization (WHO) Reproductive Health Library (RHL), and the Turning Research into Practice database (TRIP). The most recent date of search was 6 May 2019. We identified additional references by handsearching the references of articles, meta-analyses, and evidence-based guidelines retrieved from the computerised databases, on-line trial registries, and grey literature through OpenSIGLE, National Technical Information Service (NTIS), ProQuest Dissertations & Thesis Database, and Index to Theses in Great Britain and Ireland.

Selection criteria: 

Prospective or retrospective diagnostic case-control or cross-sectional studies, irrespective of publication date, format, and language, which evaluated the diagnostic accuracy of total serum bile acids (TSBA) or components of serum bile acid profile for the diagnosis of intrahepatic cholestasis of pregnancy in pregnant women of any age or ethnicity, in any clinical setting, symptomatic for pruritus.

Data collection and analysis: 

We selected studies by reading titles, abstracts, or full texts, and assessing their fulfilment of our inclusion criteria. We emailed primary authors to request missing data or individual participant data. Having extracted data from each included study, we built the two-by-two tables for each primary study and for all the index tests considered. We estimated sensitivity and specificity with their 95% confidence intervals (CI). We presented data in coupled forest plots, showing sensitivities and specificities of each study, and we plotted the studies in the Receiver Operating Characteristic (ROC) space. We performed meta-analyses adopting the hierarchical summary ROC model (HSROC) or the bivariate model to meta-analyse the data. We made indirect comparisons of the considered index tests by adding the index tests as covariates to the bivariate or HSROC models. We performed heterogeneity analysis and sensitivity analysis on studies assessing TSBA accuracy. We used Review Manager 5 (RevMan 5) and SAS statistical software, release 9.4 (SAS Institute Inc., Cary, NC, USA), to perform all statistical analyses. We used QUADAS-2 domains to assess the risk of bias of the included studies.

Main results: 

Our search yielded 5073 references, but at the end of our selection process, only 16 studies fulfilled the review inclusion criteria. Nine of these provided individual participant data. We analysed only data concerning TSBA, cholic acid (CA), glycocholic acid (GCA), chenodeoxycholic acid (CDCA), and CA/CDCA because the remaining planned index tests were assessed in few studies. Only one study had low risk of bias in all four QUADAS-2 domains. The most biased domains were the patient sampling and the reference standard domains. When considering all studies with a cut-off of 10 μmol/L, TSBA overall sensitivity ranged from 0.72 to 0.98 and specificity ranged from 0.81 to 0.97. After a sensitivity analysis excluding case-control studies, TSBA sensitivity ranged from 0.48 to 0.66 and specificity from 0.52 to 0.99. After a sensitivity analysis excluding studies in which TSBA was part of the reference standard, TSBA sensitivity ranged from 0.49 to 0.65 and specificity from 0.53 to 0.99. We found the estimates of the overall accuracy for some serum bile acid components (CA, GCA, CDCA, and CA/CDCA) to be imprecise, with the CI for sensitivity and specificity very wide or impossible to calculate. Indirect comparisons between serum bile acid profile components and TSBA were not statistically significant. None of the heterogeneity analysis performed was statistically significant, except for the timing of assessment of TSBA (onset of symptoms, peak value among multiple assessments, delivery) but without clinically relevant results. We could not analyse the diagnostic accuracy of combinations of index tests because none of the included studies carried them out, and because of the small number of included studies.

Share/Save