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

Table 7

Results showing weighted strength against possible bias risk for prospective 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 across the five factors
Harden et al. 2003 [32]	77 participants, 61.6% female, awaiting TKR in single centre setting	16 develop CRPS 1 by 1 month and 7 by 6 months after TKR; Mixed within/between designs	No mention of funding or conflict of interests	All samples assessed by the same physician	Point-biserial correlations due to small sample of CRPS
	Yes	Yes	Yes	Yes	Yes	Weak
Schürmann et al. 2000 [33]	27 participants, gender percentages not given, all with distal radial fractures in single centre setting	4 out of 27 develop CRPS 1 and two are identified as borderline	Acknowledgment is given to funder support, potential conflict of interests is not mentioned	Consensus between examiners only for oedema,; reliable Doppler perfusion monitor	Regression analysis
	Yes	Yes	Yes	Yes	No	Weak
Puchalski and Zyluk 2005 [34]	121 patients, gender percentages not given, approached with distal displaced radial fractures, the day after fixation of the fracture in single centre setting	59 refuse to participate in psychological examination; 49.5% of the sample group is available for analysis	No mention of funding or conflict of interests	Sample assessed by “we,” but clarity about authors assessment for agreement is not mentioned	A Mann-Whitney -test for determining statistical relationships
	Yes	Yes	Yes	Yes	No	Weak
Beerthuizen et al. 2012 [35]	Multicentre setting of 3 hospitals in single city, telephonic interview of 748 patients, 63.6% female, with single fracture of wrist, scaphoid, ankle, or metatarsal	596 participate (80%) and 18.1% of those who developed possible CRPS signs refused or were unable to attend	Acknowledgment is given to 2 sources of funding and neither funders are involved with design, conduct, preparation, review, or approval of the manuscript	Routine examination followed up by single experienced pain specialist clinician to confirm CRPS 1 diagnosis	A Mann-Whitney -test due to skewed distribution of variables, binary logistic regression analysis using SPSS
	No	Yes	No	Yes	No	Acceptable
Dijkstra et al. 2003 [36]	All patients, gender percentages not given, with fracture of distal radius who visit a single setting approached the day after the fracture	91 participants, 3 drop-outs; only 1 female develops CRPS 1	No funding acknowledged and hospital staff thanked for cooperation	Only researchers given as assessors for CRPS 1, no other confirmation of diagnosis	Descriptive statistics used and analysis not possible with only one CRPS 1 subject
	Yes	Yes	Yes	Yes	Yes	Weak
Dilek et al. 2012 [37]	All patients, 64% female, presenting to single setting with fractures of distal radius asked to participate in psychological assessment 2 days after cast application	74 participants with 24 drop-outs	No conflict of interests identified and no funders acknowledged, thanks given to patients who participated	No confirmation validation given for CRPS 1 diagnosis other than fulfillment of IASP criteria	Comparative statistics described with value for significance but no statistical approach described
	No	Yes	Yes	Yes	Yes	Weak
Jellad et al. 2014 [38]	All patients, 62.2% female, presenting to single setting with closed fractures of distal radius referred by orthopaedic surgeons for rehabilitation	90 participants with zero drop-outs.	No mention of funding or conflict of interests	No confirmation validation given for CRPS 1 diagnosis other than fulfillment of Veldman criteria	Logistic regression analysis using SPSS
	Yes	Yes	Yes	Yes	No	Weak
Goris et al. 2007 [39]	Multicentre setting of 3 hospitals in 2 cities, 83.3% female	114 participants with 6 drop-outs 1 year later	No benefit of any form is declared	Confirmation of diagnosis using 2 criteria as well as assessment by 2 investigators and blinding to results at a year’s follow-up	Mann-Whitney -test for paired data, Kruskal-Wallis test for nonparametric data, Gunn’s multiple comparison test with regression coefficient for change of regional inflammation score
	No	No	No	No	No	Strong
Gradl and Schürmann 2005 [40]	All consecutive patients, 60% female, who develop CRPS 1 following trauma in single centre setting	10 participants	No mention of funding or conflict of interests	Confirmation of diagnosis using Harden/Bruehl criteria by 2 independent surgeons and pain specialist investigators and blinding to results at a year’s follow-up	Repeated measures every 10 days for 3 months with result of measures graphically presented but no statistical approach described
	Yes	Yes	Yes	No	Yes	Weak
Moseley et al. 2014 [41]	1661 patients, 51.5% female presenting to multicentre (3) hospital settings with fractures of distal radius not needing surgical fixation; no city(s) mentioned	1549 participants, 21 drop-outs due to administrative error, only mentioned as percentages	No mention of funding or conflict of interests	Routine examination by single experienced pain specialist clinician to confirm CRPS 1 diagnosis made by “standard criteria” listed in an appendix	A predictive model was developed using logistic regression, likelihood ratio test, with bootstrap sampling and goodness of fit with Hosmer-Lemeshow test
	No	No	Yes	Yes	No	Acceptable