Table 2. Statistical analysis of all cases for Enneking functional score.

			Summary statistic		Univariate		Multivariate: initial modelP<0.05 in univariate		Multivariate: final modelP<0.05 in step-down method
Factor	Category	N	Average	SD	95% CI	P value	95% CI	P value	95% CI	P value
Gender	Male	30	22.1	4.5	–	[0.031*]	–	[0.292]
	Female	16	19.2	3.8	(−5.616, −0.276)	0.031*	(−3.674, 1.140)	0.292
Chemotherapy	No	26	22.3	5.0	–	[0.036*]	–	[0.108]
	Yes	20	19.6	3.1	(−5.330, −0.185)	0.036*	(−4.101, 0.425)	0.108
Resection range	Total scapulectomy	25	19.0	3.7	–	[0.003**]	–	[0.118]	–	[0.006**]
	Acromion preserved	7	22.6	3.2	(0.245, 6.898)	0.036*	(0.074, 6.279)	0.045*	(0.503, 6.593)	0.024*
	Glenoid preserved	3	21.3	3.1	(−2.420, 7.087)	0.327	(−2.236, 6.409)	0.334	(−1.799, 6.881)	0.243
	Both of acromion and glenoid preserved	9	24.7	4.7	(2.643, 8.691)	<0.001***	(0.517, 6.970)	0.024*	(2.615, 8.137)	<0.001***
	Resection of lower half	2	26.0	5.7	(1.283, 12.717)	0.018*	(−4.650, 9.931)	0.467	(−5.511, 8.696)	0.652
Resected nerve	No	34	22.0	4.3	–	[0.017*]	–	[0.246]
	Axillary	12	18.5	4.1	(−6.390, −0.669)	0.017*	(−4.061, 1.076)	0.246
Follow-up term (mons)	<20	13	17.8	3.0	–	[0.002**]	–	[0.039*]	–	[0.005**]
	≥20<70	17	22.8	4.8	(2.038, 7.799)	0.001**	(0.357, 5.804)	0.028*	(1.182, 6.362)	0.005**
	≥70	15	21.5	3.3	(0.658, 6.583)	0.018*	(1.016, 6.463)	0.009**	(1.451, 6.807)	0.003**
	Unknown	1	30.0		(4.040, 20.268)	0.004**	(−2.096, 19.365)	0.111	(2.085, 22.174)	0.019*

Multivariate analysis was performed using seventeen factors of the patient’s background to determine which influence Enneking’s functional score or active range of motion for all cases and for total scapulectomy cases, separately. The amount of remaining bone influenced the Enneking functional score, which means that preserving the glenoid or the acromion lead to better function compared to total scapulectomy. However, there was no significant evidence that reconstruction improved total functional outcome.