Potential Risk Factors for the Onset of Complex Regional Pain Syndrome Type 1: A Systematic Literature Review

Table 8

Results showing weighted strength against possible bias risk for retrospective studies’ analyses with risk rating in bold italics.

Authors	Population sample selection bias risk	Study design bias risk	Funding provision bias risk	Detection bias risk	Measurement bias risk	Weighted strength
Allen et al. 1999 [42]	Medical records of all consecutive CRPS 1 patients, 70% female, who were referred to a single setting multidiscipline pain centre	134 consecutive CRPS 1 patients	No mention of funding or conflict of interests	Independent review of medical records	Result of measures listed with values for significance but no statistical approach described
	No	Yes	Yes	No	Yes	Weak
de Mos et al. 2007 [5]	Electronic medical records of one country, population gender mix not given	Estimation of CRPS 1 incidence from data of 600 000 patients; control groups are those who do not develop CRPS 1	Acknowledgment of funding provided through TREND, a Dutch government research grant; no mention of conflict of interests	Potentially overstrict retrospective application of 3 sets of criteria: Veldman, IASP, and Harden/Bruehl, to electronic records by independent reviewers with statistics for agreement	Poisson distribution, logistic regression, chi-square, and Student’s -tests using SPSS software
	No	No	No	No	No	Strong
Sandroni et al. 2003 [4]	Electronic medical records of one region, population gender mix not given	Estimation of CRPS 1 incidence from data of 70 745 patients	Acknowledgment of funding provided through two sources, no conflict of interests mentioned	Potentially lenient retrospective application of IASP criteria to electronic records by one reviewer with 10% random independent assessment and 93.4% agreement	Chi-square, Fisher’s exact, and Wilcoxon rank sum tests with no software described
	No	No	No	Yes	No	Acceptable
Duman et al. 2007 [43]	Medical records of two tertiary military hospitals in one country, all male	168 CRPS 1 patients	No mention of funding or conflict of interests	Author review with no independent assessment using IASP criteria and three-phase bone scan	Descriptive statistics only using SPSS
	Yes	Yes	Yes	Yes	Yes	Weak
van Rijn et al. 2007 [22]	Medical records of all CRPS patients, 86.5% female, who were referred to a single setting movement disorder centre	121 CRPS 1 patients	Acknowledgment of funding provided through TREND, a Dutch government research grant; no mention of conflict of interests	Author review with no independent assessment using IASP criteria	A multivariate analysis using Cox’s proportional hazards model, Mann-Whitney -test, chi-square tests, and Student’s -tests using SPSS software
	No	Yes	No	Yes	No	Acceptable
Anderson and Fallat 1999 [44]	Medical records of all CRPS patients, 61% female, seen in a single setting foot and ankle trauma clinic	33 CRPS 1 patients	No funding mentioned and acknowledgment of assistance with statistical analysis and illustrations	CRPS 1 diagnosis confirmed in records by medical specialist at pain management centre	Regression analysis, unpaired -test, paired -test, and Pearson’s correlation using SPSS
	Yes	Yes	Yes	Yes	No	Weak