Risk factors for revision after shoulder arthroplasty

Figures & data

Table 1. General characteristics and diseases leading to shoulder replacement

Diagnosis ^a	Total no. of patients	Hemiprosthesis	Reversed TSR	Neer-type TSR
Total	1,820	1,526	225	69
Age (mean, SD)	69 (12)	69 (13)	69 (11)	68 (11)
Sex (% women)	78%	78%	83%	69%
Diagnosis
Osteoarthritis	427	338	50	39
Rheumatoid arthritis	569	439	118	12
Acute fracture	426	422	4	–
Fracture sequelae	350	303	34	13
Ankylosing spondylitis	16	14	1	1
Psoriatic arthritis	18	15	3	–
Rotator cuff /ligament damage	23	10	13	–
Other ^b	64	48	11	5

^a More than one diagnosis was allowed.

^b Avascular necrosis, sequelae after infection, hemophilia, cancer, sequelae after luxation, SLE, chondrocalcinosis, sequelae after frozen shoulder, synovial chondromatosis, and unknown.

Table 2. Prosthesis brands

Prosthesis brand	HP	Neertype TSR	Reversed TSR	Hospitals using the HP	Hospitals with > 10 HPs ^a	Hospitals using the TSR	Hospitals with > 10 TSRs ^a
Bio-Modular (Biomet)	604	48		26	14 (7)	9	2
Global (DePuy)	261	1		21	8 (3)	1	0
Delta III (DePuy)			225			17	6 (3)
Copeland (Biomet)	121	4		15	3 (2)	3	0
Nottingham (Biomet)	109	13		4	2 (2)	1	1
Global Advantage (DePuy)	105	1		15	1 (1)	1	0
Global Fx(DePuy)	119			17	0
Delta I (Medinov)	58			7	2 (1)
Scan Shoulder (Mitab)	56			7	2
Neer II (3M Healthcare)	40			6	1
Modular (3M Healthcare)	33			1	1 (1)
Monospherical (Howmedica)	13	1		3	0	1	0
Bigliani Flatow (Zimmer)	6			2	0
	1,525 ^b	68 ^b	225

HP: hemiprosthesis; TSR: total shoulder replacement.

^a Number in parentheses: hospitals with > 30 procedures

^b For one TSR and one HP, the brand of prosthesis was not given.

Figure 1. Shoulder arthroplasties in 4 major diagnostic groups from 1994 through 2005. IA: inflammatory arthritis; OA: osteoarthritis.

Table 3. Primary shoulder arthroplasties per annum. Overall incidence (per 100,000 inhabitants), total annual number, and annual number according to prosthesis type

Year	Overall incidence	All primary operations	Hemiprostheses	Total prostheses
1994	2,80	121	93	28
1995	2,67	116	87	29
1996	2,77	121	94	27
1997	3,37	148	133	15
1998	2,90	128	118	10
1999	3,55	158	135	23
2000	3,13	140	114	26
2001	3,24	146	130	16
2002	3,34	151	122	29
2003	3,98	181	153	28
2004	4,33	198	168	30
2005	4,71	217	184	33
Total		1820	1526	294
p-value ^a	< 0.001	< 0.001	< 0.00	0.31

^a p-value calculated using Poisson regression analysis based on annual population rates for the Norwegian population.

Figure 2. Kaplan-Meier failure curve (with 95% CI shown in color) for all hemiprostheses, resurfacing hemiprostheses, reversed total shoulder replacement, and Neer-type total shoulder replacement.

Figure 3. A. Kaplan-Meier failure curves for different hemiprosthesis brands. Groups with less than 20 patients were excluded. B. Kaplan-Meier failure curves according to shoulder disease.

Table 4a. Cox regression analysis for revision after hemiarthroplasty, according to potential explanatory factors. 5- and 10-year failure rates estimated using the Kaplan-Meier method

	No.	No. of revisions	F5 (%)	95% CI	F10 (%)	95% CI	RR	95% CI	p-value
All	1,470 ^a	74	6	5–7	8	6–10
Age
0–69 years	664	49	9	7–12	12	9–15	1
≥ 70 years	806	25	4	3–5	4	3–5	0.47	0.28–0.77	0.003
Sex
Men	312	14	6	3–10	8	4–12	1
Women	1158	60	6	5–8	8	6–10	1.5	0.82–2.7	0.2
Year of surgery							1.0	0.92–1.1	1
Diagnosis
Acute fracture	422	9	1	0–2	3	1–4	1
IA ^b	440	28	6	3–8	9	6–12	1.8	0.80–3.9	0.1
OA ^b	311	14	6	3–9	9	4–14	2.2	0.95–5.1	0.07
Fracture
sequelae	297	23	10	6–14	10	6–14	3.3	1.5–7.2	0.002

^a61 ca ses were not included due to missing values or due to the fact that the patient had a shoulder disease other than the 4 included in this analysis.

^b OA: osteoarthritis; IA: inflammatory arthritis including RA, psoriatic arthritis, and ankylosing spondylitis.

F5 and F10: unadjusted Kaplan-Meier 5-year and 10-year failure rates.

RR: relative risk of revision derived from the Cox model.

Table 4b. Cox regression survival analysis for revision after reversed TSR, according to potential explanatory factors. 5- and 10-year failure rates estimated using the Kaplan-Meier method

	No.	No. of	F5 (%) revisions	95% CI	F10 (%)	95% CI	RR	95% CI	p-value
All		201 ^a	22	10	85–95	22	67–90
Age
0–69 years	93	13	11	4–18	26	11–41	1
≥ 70 years	108	9	8	3–14	12	3–20	0.82	0.32–2.12	0.7
Sex
Men	32	9	30	11–50	51	23–80	1
Women	169	13	7	2–11	15	4–26	0.26	0.11–0.63	0.003
Year of surgery						1.09	0.93–1.3	0.3
Diagnosis
IA ^b	120	14	9	3–14	20	8–32	1
OA ^b	47	6	14	1–28	36	0–73	1.49	0.52–4.3	0.5
Fracture sequelae	34	2	10	0–23	10	0–23	0.64	0.14–2.9	0.6

^a 24 cases were not included due to missing values or due to the fact that the patient had a shoulder disease other than the 3 included in this analysis.

^b OA: osteoarthritis; IA: inflammatory arthritis including RA, psoriatic arthritis, and ankylosing spondylitis.

F5 and F10: unadjusted Kaplan-Meier 5-year and 10-year failure rates.

RR: relative risk of revision derived from the Cox model.

Table 5. Causes of revision, according to type of prosthesis

Cause ^a	All (n = 1,820)	Hemiprosthesis (n = 1,526)	TSR reversed (n = 225)
Aseptic loosening of glenoid component	17	–	13
humeral component	11	6	5
Luxation	21	11	4
Instability	9	6	2
Deep infection	11	5	5
Fracture	3	2	1
Pain	34	32	2
Failure of function	3	3	–
Other ^b	26	20	4

^a More than one cause may have been given, but pain was included only when it was given as the sole cause of revision.

^b Including incorrect axis, broken prosthesis component, component too small or too large, subluxation, and other not specified.

Table 6a. Risk of revision due to specific causes, for hemiprostheses (HPs), according to shoulder disease and fixation. The results were obtained using an exact regression model

	Aseptic loosening			Pain ^d			Deep infection			Instability			Luxation of prosthesis
	n ^a	RR ^b (95% CI)	p	n ^a	RR ^b (95% CI)	p	n ^a	RR ^b (95% CI)	p	n ^a	RR ^b (95% CI)	p	n ^a	RR ^b (95% CI)	p
Shoulder disease
IA ^c (439)	4	1		15	1	1	1			0	1		0	1
OA ^c (311)	2	1.1	1.0	5	1.0	1.0	2	2.5	0.87	1	4.2	0.38	2	2.8	0.4
		(0.09–9.0)			(0.28–3.1)			(0.12–156)			(0.11–∞)			(0.22–∞)
Sequelae of fracture (296)	1	0.53	0.99	8	1.5	0.52	1	1.4	1.0	3	11.4	0.04	5	8.8	0.04
		(0.01–5.7)			(0.53–3.8)			(0.02–119)			(1.1–∞)			(1.1–∞)
Acute fracture (421)	0	0.28	0.27	0	0.12	0.02	0	1.1	1.0	1	1.8	0.72	2	1.7	0.7
		(0.0–2.5)			(0.0–0.80)			(0.0–41)			(0.05–∞)			(0.13–∞)
Fixation
Cemented (944)	5	1		10	1		2	1		3	1			6	1
Uncemented (523)	2	0.28	0.24	18	1.4	0.56	2	1.2	1.0	2	0.91	1.0	3	1.4	0.9
		(0.03–1.8)			(0.59–3.4)			(0.08–17)			(0.07–9.9)			(0.21–7.6)

In the analyses, some cases were excluded due to missing values (i.e. patients with diagnoses other than the major ones).

^a Number of revisions due to each cause.

^b Adjusted for age and sex. Revision due to pain was significantly less frequent in patients who were 70 years or older compared to those who were younger (RR = 0.25, p = 0.005).

^c OA: osteoarthritis; IA: inflammatory arthritis including RA, psoriatic arthritis, and ankylosing spondylitis.

^d Pain was registered only when it was given as the only cause of revision.

Table 6b. Risk of revision due to specific causes, after reversed TSR, according to shoulder disease and fixation. The results were obtained using an exact regression model

	Aseptic loosening			Pain ^d			Deep infection			Instability			Luxation of prosthesis
	n ^a	RR ^b (95% CI)	p	n ^a	RR ^b (95% CI)	p	n ^a	RR ^b (95% CI)	p	n ^a	RR ^b (95% CI)	p	n ^a	RR ^b (95% CI)	p
Shoulder disease
IA ^d (120)	9	1		1	1		2	1		1	1		2	1
OA ^d (47)	3	1.6	0.8	1	2.2	1	0	2.0	1	1	1.3	1.0	2	2.1	0.9
		(023–8.4)			(0.03–164)			(0.0–26)			(0.02–106)			(0.13–38)
Sequelae of fracture (296)	1	0.62	1	0	4.3	1	1	1.8	1	0	6.5	1.0	0	1.4	1
		(0.01–5.7)			(0.0–166)			(0.03–35)			(0.0–252)			(0.00–18)
Fixation
Cemented (61)	1	1		1	1		1	1		0	1		2	1
Uncemented (264)	13	1.5	1	1	0.08	0.4	4	0.15	1	2	0.68	1.0	2	0.31	0.5
		(0.19–68)			(0.0007–9.4)			(0.0004–∞)			(0.05–∞)			(0.02–4.6)

In the analyses, some cases were excluded due to missing values (i.e. patients with diagnoses other than the 3 major ones).

^a Number of revisions due to each cause.

^b Adjusted for age and sex. Revision due to deep infection was significantly less frequent in women than in men (RR = 0.06, p = 0.01).

^c Loosening of humeral or glenoid component, or both.

^d OA: osteoarthritis; IA: inflammatory arthritis including RA, psoriatic arthritis, and ankylosing spondylitis.

^e Pain was registered only when it was given as the only cause of revision.