|
Correlation of biochemical response to interferon alfa with histological improvement in hepatitis C: a meta-analysis of diagnostic test characteristics |
Bonis P A, Ioannidis J P, Cappelleri J C, Kaplan M M, Lau J |
|
|
Authors' objectives To determine the extent to which biochemical measures reflect histological outcomes in the treatment of hepatitis C with interferon (IFN)-alpha.
Searching MEDLINE was searched from 1966 to December 1995 using indexing terms noted from articles and the following keywords: 'hepatitis C', 'non-a non-b hepatitis' and 'interferons'. In addition, the bibliographies of retrieved papers were examined.
Study selection Study designs of evaluations included in the reviewStudies where biopsies were performed only for biochemical responders were excluded.
Specific interventions included in the reviewStudies using alanine aminotransferase (ALT) or detection of viral RNA as index tests were eligible for inclusion. Biochemical responders were defined as those whose ALT was normal at the time of repeat biopsy. Sustained responders were defined as those with a normal ALT for 6 months or longer after discontinuation of the treatment. RNA response was defined as undetectable serum RNA at the time of post-treatment biopsy, and detectable RNA pre- treatment.
Reference standard test against which the new test was comparedOnly studies using liver biopsy performed before and after therapy as the reference standard were eligible for inclusion. Histological improvement was defined differently across the studies, as the Knodell score (see Other Publications of Related Interest no.1), modified histological activity score, and conventional histological assessment.
Participants included in the reviewStudies of patients with chronic hepatitis C, or non-a non-b hepatitis, who had received IFN-alpha, were included. Studies of patients with other forms of liver disease were excluded.
Outcomes assessed in the reviewNo inclusion criteria relating to the outcome measures were specified. The outcomes calculated in the review were the sensitivity and specificity of ALT. The sensitivity of ALT was defined as the percentage of patients with a normal ALT level at the time of the second liver biopsy among those who had improved hepatic histology. The specificity of ALT was defined as the percentage of patients with an abnormal ALT level among those whose liver histology either worsened or remained unchanged after therapy.
How were decisions on the relevance of primary studies made?Non-English language articles were reviewed with the help of translators. No other details of the process by which the studies were selected, were given.
Assessment of study quality The method presented by Mulrow et al. (see Other Publications of Related Interest no.2) was employed. In addition, there was an assessment of whether investigators blinded to the clinical status and biochemical responses of the participants conducted the histological scoring of liver biopsy specimens, whether the scoring was performed in duplicate, and whether the details of the scoring system were described adequately. Additional criteria were the assessment of the percentage of patients who underwent a second liver biopsy, and the reasons why a second biopsy was not performed. The authors did not state how the papers were assessed for validity, or how many reviewers performed the validity assessment.
Data extraction Two reviewers extracted the data independently. Any discrepancies were resolved by recourse to a third reviewer.
Methods of synthesis How were the studies combined?Summary receiver operating characteristic (ROC) curves were estimated for ALT; these were estimated separately for the different definitions of liver histology. Pooled estimates of sensitivity and specificity were calculated using a random-effects model, where the studies were weighted by the inverse variance. For studies providing both ALT and RNA data, the correlation between these was examined. The correlations were standardised using Fischer's Z transformation.
How were differences between studies investigated?The chi-squared test was used to assess between-study heterogeneity. The data were analysed separately for strict and less stringent criteria for histological improvement. The threshold effect was examined using a logit transformation of the summary ROC model and regression analysis.
Results of the review There were 15 studies (n=1,546) for the quantitative assessment and 42 studies (number of participants not reported) for the qualitative assessment.
Qualitative assessment.
In 27 of the 42 studies it was stated that there was discordance between histological and biochemical outcomes, or a trend toward improvement among patients without a normalised ALT level, in response to IFN.
Quantitative assessment.
Significant between-study heterogeneity was present. The threshold effect was not significant.
Strict definitions of histological improvement (9 studies, n=201): among participants who underwent a second liver biopsy at the end of treatment, 28% (95% confidence interval, CI: 17, 43) had improved histology, whilst for 5% (95% CI: 2, 10), histology had deteriorated. The sensitivity and specificity of ALT for determining histological change were 70% (95% CI: 56, 81) and 66% (95% CI: 56, 75), respectively. As many as 17% (95% CI: 9, 30) of the patients with an abnormal ALT level at the end of therapy may have improved histologically. The regression analysis showed that the normalised ALT among patients with improved liver histology was about 4.8 (95% CI: 3.1, 7.4) times more likely than a normalised ALT among patients without improved liver histology.
Less stringent definition of histological improvement (12 studies, n=437): among participants who underwent a second liver biopsy at the end of treatment, 62% (95% CI: 51, 72) had improved histology, whilst for 13% (95% CI: 6, 25), histology had deteriorated. The sensitivity and specificity of ALT for determining histological change were 55% (95% CI: 44, 65) and 75% (95% CI: 67, 81), respectively. As many as 51% (95% CI: 38, 64) of the patients with an abnormal ALT level at the end of therapy may have improved histologically. The regression analysis showed that the normalised ALT among patients with improved liver histology was about 4.3 (95% CI: 2.8, 6.5) times more likely than a normalised ALT among patients without improved liver histology. Statistically- significant variability in sensitivity rates (P<0.025) and negative predictive value rates (P<0.0025) were found between studies when the less stringent definitions of histological response were considered.
Five studies reported RNA response. For the strict definition of histological improvement, the sensitivity ranged from 0 to 100% and the specificity from 41 to 100%. For the less stringent definition, the sensitivity ranged from 0 to 100% and the specificity from 25 to 100%. For studies reporting ALT and RNA results, the correlation coefficients ranged from 0.23 to 0.80.
Authors' conclusions A substantial number of patients may improve histologically after IFN therapy. The significance of histological changes observed after IFN therapy must be weighed against the limitations of liver biopsy and the uncertain natural history of hepatitis C. Nevertheless, the ALT level does not always reflect liver histology accurately after IFN-alpha treatment and may underestimate histological improvement.
CRD commentary The research question, the inclusion and exclusion criteria for the primary studies, the criteria for the validity assessment, and the data pooling were all clearly presented in this review. However, the pooling of heterogeneous sensitivity and specificity values is questionable without further investigation of potential sources of heterogeneity. More details of the review process (i.e. how the studies were selected, the number of reviewers involved, and how any disagreements were resolved) would have been useful. Some details of the primary studies were tabulated, but more information about their design would have been of interest. It is possible that eligible studies may have been omitted from the review as the literature search involved only one database, with no strategy to locate unpublished material.
It should be noted that the findings of this review cannot be used to support the effectiveness of IFN-alpha in hepatitis C, as the objective was to assess the predictive value of ALT with reference to histological improvement.
Implications of the review for practice and research Practice: The authors stated that the available data do not support the use of repeated liver biopsies for assessing histological changes after IFN treatment.
Research: The authors stated that there is a need for further studies to better characterise histological responses to IFN treatment, to understand which patients are most likely to improve histologically after treatment, and to understand the impact of these changes on the natural history of the disease. They further stated that there is a great need for better understanding of the utility of each diagnostic test and its implications for treatment of the disease.
Funding National Institutes of Health, grant numbers T32 DK07701-02, T32 A107389, RO1 HS07782; National Institute of Diabetes and Digestive and Kidney Diseases, Gastrointestinal Research on Absorptive and Secretory Processes Digestive Disease Center, U.S. Public Health Service, grant number P30 DK34928.
Bibliographic details Bonis P A, Ioannidis J P, Cappelleri J C, Kaplan M M, Lau J. Correlation of biochemical response to interferon alfa with histological improvement in hepatitis C: a meta-analysis of diagnostic test characteristics. Hepatology 1997; 26(4): 1035-1044 Other publications of related interest 1. Knodell RG, Ishak KG, Black WC, Chen TS, Craig R, Kaplowitz N, et al. Formulation and application of a numerical scoring system for assessing histological activity in asymptomatic chronic active hepatitis. Hepatology 1981;1:431-5. 2. Mulrow CD, Linn WD, Gaul MK, Pugh JA. Assessing the quality of a diagnostic test evaluation. J Gen Intern Med 1989;4:288-95.
Indexing Status Subject indexing assigned by NLM MeSH Alanine Transaminase /blood; Hepatitis C /metabolism /pathology /therapy; Humans; Interferon-alpha /therapeutic use; Liver /pathology; Sensitivity and Specificity AccessionNumber 11997001297 Date bibliographic record published 30/04/2004 Date abstract record published 30/04/2004 Record Status This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn. |
|
|
|