Search Update

We updated the search to December 2010 restricting to MEDLINE (1950 to December Week 3 2010), the Cochrane Library (2010 4), EMBASE (1980 to 2010 Week 52), and Ovid Healthstar (1966 to December 2010). Title and abstract of 100 new records were screened per protocol. Thirteen records passed to full text screening level. Four records were excluded because they were editorial or reviews, and five because of miscellaneous reasons [6107, 6147, 6152, 6250, 6168, 6154, 6008, 6011, 6136]. Of the four new studies included, one lacked analyzable data [6167]. Two studies contributed to KQ1a and 1b [6017, 6140], and one fair quality cohort contributed to KQ3c [6151].

Methods

As the data were sparse, we made the following modifications to the previous methodology to examine robustness of conclusions.

• for both diagnostic accuracy (KQ1c) and investigation of genotypic association(KQ3c), we undertook additional (post hoc) meta-analyses in which all studies were pooled as long as they genotyped at the very least for all ethnicity specific mutations with known prevalence >1%. That is, TPMT*3A for Caucasians, and TPMT*3C for Asians—there were zero studies on participants of African descent.
• When appropriate, sensitivity and specificity estimates were pooled by first transforming proportions into the Freeman-Tukey variant of the arcsine square root transformed proportion [6256]. The pooled proportion was calculated as the back-transformation of the weighted mean of the transformed proportions. Data were pooled using a fixed effects inverse variance weighted average. We revised previous primary meta-analyses and conducted additional meta-analyses where applicable.
• We pooled odds ratios using the fixed effects Mantel-Haenszel method without continuity correction [6254].

Results

Revised Conclusions

There is a dearth of good quality primary research addressing the comparative effectiveness of TPMT pretesting and test performance of TPMT genotyping with respect to enzymatic activity testing.

Evidence of limited quality indicates that the estimates of sensitivity of genotyping are imprecise, despite near perfect specificity, to identify subnormal enzymatic activities. There is currently insufficient evidence addressing the utility of TPMT testing prior to commencement of thiopurine therapy in comparison with routine blood count monitoring. It also remains unclear whether pre-testing guides appropriate prescribing. Indirect evidence confirms previously known strong associations between low levels of TPMT enzymatic activity and the presence of TPMT allelic polymorphisms status with thiopurine related leukopenia. This was reflected in significant associations between low levels of enzymatic activity and myelotoxicity.