Original Research

ABSTRACT

Use of mesenchymal stem cells from bone marrow has gained significant popularity. The iliac crest has been determined to be an effective site for harvesting mesenchymal stem cells. Review of the literature reveals that multiple techniques are used to harvest bone marrow aspirate from the iliac crest, but the descriptions are based on the experience of various authors as opposed to studied anatomy. A safe, reliable, and reproducible method for aspiration has yet to be studied and described. We hypothesized that there would be an ideal angle and distance for aspiration that would be the safest, most consistent, and most reliable. Using magnetic resonance imaging (MRI), we reviewed 26 total lumbar spine MRI scans (13 males, 13 females) and found that an angle of 24° should be used when entering the most medial aspect of the posterior superior iliac spine (PSIS) and that this angle did not differ between the sexes. The distance that the trocar can advance after entry before hitting the anterior ilium wall varied significantly between males and females, being 7.53 cm in males and 6.74 cm in females. In addition, the size of the PSIS table was significantly different between males and females (1.20 cm and 0.96 cm, respectively). No other significant differences in the measurements gathered were found. Using the data gleaned from this study, we developed an aspiration technique. This method uses ultrasound to determine the location of the PSIS and the entry point on the PSIS. This contrasts with most techniques that use landmark palpation, which is known to be unreliable and inaccurate. The described technique for aspiration from the PSIS is safe, reliable, reproducible, and substantiated by data.

The iliac crest is an effective site for harvesting bone marrow stem cells. It allows for easy access and is superficial in most individuals, allowing for a relatively quick and simple procedure. Use of mesenchymal stem cells (MSCs) for treatment of orthopedic injuries has grown recently. Whereas overall use has increased, review of the literature reveals very few techniques for iliac crest aspiration,¹ but these are not based on anatomic relationships or studies. Hernigou and colleagues^2,3 attempted to quantitatively evaluate potential “sectors” allowing for safe aspiration using cadaver and computed tomographic reconstruction imaging. We used magnetic resonance imaging (MRI) to analyze aspiration parameters. Owing to the ilium’s anatomy, improper positioning or aspiration technique during aspiration can result in serious injury.^2,4-6 We hypothesized that there is an ideal angle and positioning for bone marrow aspiration from the posterior superior iliac spine (PSIS) that is safe, consistent, and reproducible. Although most aspiration techniques use landmark palpation, this is unreliable and inaccurate, especially when compared with ultrasound-guided injections^7-16 and procedures.^9,12,17-19 We describe our technique using ultrasound to visualize patient anatomy and accurately determine anatomic entry with the trocar.

METHODS

MRI scans of 26 patients (13 males, 13 females) were reviewed to determine average angles and distances. Axial T2-weighted views of the lumbar spine were used in all analyses. The sacroiliac (SI) joint angle was defined as the angle formed between the vector through the midline of the pelvis and the vector that is parallel to the SI joint. The approach angle was defined as the angle formed between the vector of the most medial aspect of the PSIS through the ilium to the anterior wall and the vector through the midline of the pelvis (Figure 1). 

The distance, in centimeters, of the PSIS to the anterior ilium wall was measured to determine the maximum distance the trocar can advance without puncturing the anterior ilium wall. The perpendicular distance from the PSIS table to the anterior aspect of the ilium was measured to determine the average depth the trocar could advance if the clinician entered it perpendicular to the patient’s coronal plane (Figure 2). 

The PSIS table width was defined as the width, in centimeters, of the PSIS posteriorly. The minimum width, in centimeters, of the ilium was measured at the thinnest region of the ilium to determine the degree of variation if the trocar were entered too far laterally and advanced to the anterior ilium wall (Figure 3). 

Means and standard deviations were calculated by sex and as a total. Student’s t test was used to determine statistical significance (P < .05).

Continue to: For the 13 males, the mean SI joint...

 

RESULTS

The results are reported in the Table. 

Table. Measurements of Patients Taken on Axial T2-Weighted Views of Lumbosacral MRI Scans^a

Patient	SI Joint Angle (°)	Approach Angle (°)	PSIS Table Width (cm)	PSIS to Anterior Ilium Wall (cm)	Perpendicular Distance PSIS to Anterior Joint (cm)	Post Ilium Wall to SI Joint Width (cm)
Males
1	28.80	19.50	1.24	8.80	4.16	1.52
2	31.80	27.60	1.70	7.89	3.49	1.02
3	33.70	27.70	1.12	8.14	3.15	1.28
4	23.70	26.40	0.95	6.66	3.22	0.65
5	35.90	28.40	0.84	7.60	2.57	0.95
6	33.80	29.30	1.20	7.73	2.34	0.90
7	30.30	21.20	1.36	8.44	3.95	1.18
8	34.50	20.40	1.53	7.08	3.98	1.56
9	28.70	24.00	1.34	8.19	3.51	1.31
10	22.40	20.10	1.37	7.30	3.87	1.28
11	33.60	20.80	0.88	6.43	3.26	0.94
12	48.50	31.00	1.15	6.69	2.97	1.38
13	20.20	20.90	0.94	6.95	3.79	1.05
Averages	31.22	24.41	1.20	7.53	3.40	1.16
Standard Deviation	7.18	4.11	0.26	0.75	0.56	0.26
Females
14	22.80	23.20	1.54	7.21	3.45	1.39
15	33.30	21.40	1.09	7.26	3.57	0.98
16	19.70	15.60	0.78	8.32	3.76	0.86
17	17.50	15.60	0.61	7.57	3.37	1.03
18	48.20	26.60	0.94	6.62	3.16	0.71
19	38.20	28.30	0.90	6.32	2.23	0.91
20	44.50	31.70	0.99	6.19	3.06	0.76
21	24.10	18.00	0.92	6.99	3.23	0.71
22	17.20	14.80	0.81	6.00	2.81	1.13
23	42.00	38.50	1.00	5.33	2.47	1.42
24	32.00	25.50	0.98	6.01	2.79	1.21
25	24.70	24.80	0.87	6.09	2.79	1.02
26	19.80	22.30	1.04	7.71	2.37	1.36
Averages	29.54	23.56	0.96	6.74	3.00	1.04
Standard Deviation	10.84	6.88	0.21	0.85	0.48	0.25
All patients Averages	30.38	23.98	1.08	7.14	3.20	1.10
Standard Deviation	9.05	5.57	0.26	0.88	0.55	0.26

^aStatistical significance is denoted as P < .02.

Abbreviations: MRI, magnetic resonance imaging; PSIS, posterior iliac spine; SI, sacroiliac.

For the 13 males, the mean SI joint angle was 31.22° ± 7.18° (range, 20.20° to 48.50°). The mean approach angle was 24.41° ± 4.11° (range, 19.50° to 31.00°). The mean PSIS table width was 1.20 cm ± 0.26 cm (range, 0.84 cm to 1.70 cm). The mean distance from the PSIS to the anterior ilium wall was 7.53 cm ± 0.75 cm (range, 6.43 cm to 8.80 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.40 cm ± 0.56 cm (range, 2.34 cm to 4.16 cm). The mean minimum width of the ilium to the SI joint was 1.16 cm ± 0.26 cm (range, 0.65 cm to 1.56 cm).

For the 13 females, the mean SI joint angle was 29.54° ± 10.84° (range, 17.20° to 48.20°). The mean approach angle was 23.56° ± 6.88° (range, 14.80° to 38.50°). The mean PSIS table width was 0.96 cm ± 0.21 cm (range, 0.61 cm to 1.54 cm). The mean distance from the PSIS to the anterior ilium wall was 6.74 cm ± 0.85 cm (range, 5.33 cm to 8.32 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.00 cm ± 0.48 cm (range, 2.23 cm to 3.76 cm). The mean minimum width of the ilium to the SI joint was 1.04 cm ± 0.25 cm (range, 0.71 cm to 1.42 cm).

For the 26 total patients, the mean SI joint angle was 30.38° ± 9.05° (range, 17.20° to 48.50°). The mean approach angle was 23.98° ± 5.57° (range, 14.80° to 38.50°). The mean PSIS table width was 1.08 cm ± 0.26 cm (range, 0.61 cm to 1.70 cm). The mean distance from the PSIS to the anterior ilium wall was 7.14 cm ± 0.88 cm (range, 5.33 cm to 8.80 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.20 cm ± 0.55 cm (range, 2.23 cm to 4.16 cm). The mean minimum width of the ilium to the SI joint was 1.10 cm ± 0.26 cm (range, 0.65 cm to 1.56 cm).

There was a statistically significant difference between the male and female groups for the maximum distance the trocar can be advanced from the PSIS to the anterior ilium wall (P < .02), and a statistically significant difference for the PSIS table width (P < .02). There were no significant differences between the male and female groups for the approach angle, the SI joint angle, the perpendicular distance from the PSIS to the anterior ilium, and the minimum width of the ilium to the SI joint.

Continue to: The patient is brought to the procedure...

 

TECHNIQUE: ILIAC CREST (PSIS) BONE MARROW ASPIRATION

The patient is brought to the procedure room and placed in a prone position. The donor site is prepared and draped in the usual sterile manner. Ultrasound is used to identify the median sacral crest in a short-axis view. The probe is then moved laterally to identify the PSIS (Figures 4A, 4B). 

The probe can be moved superiorly and inferiorly to determine the most prominent and central portion of the PSIS. The SI joint and ilium can also be visualized if needed.

The crosshairs on the ultrasound probe are used to mark the center lines of each plane. The central point marks the location of the PSIS. Alternatively, an in-plane technique can be used to place a spinal needle on the exact entry point on the PSIS. Once the PSIS and entry point are identified, the site is blocked with 10 mL of 0.5% ropivacaine.

Prior to introduction of the trocar, all instrumentation is primed with heparin and syringes are prepped with anticoagulant citrate dextrose solution, solution A. A stab incision is made at the site. The trocar is placed at the entry point, which should be centered in a superior-inferior plane and at the most medial point of the PSIS. Starting with the trocar vertical, the trocar is angled laterally 24° by dropping the hand medially toward the midline. No angulation cephalad or caudad is necessary, but cephalad must be avoided so as not to skive superiorly. This angle, which is recommended for both males and females, allows for the greatest distance the trocar can travel in bone before hitting the anterior ilium wall. A standard deviation of 5.57° is present, which should be considered. Steady pressure should be applied with a slight twisting motion on the PSIS. If advancement of the trocar is too difficult, a mallet or drill can be used to assist in penetration.

With the trocar advanced into the bone 1 cm, the trocar needle is removed while the cannula remains in place. The syringe is attached to the top of the cannula. The syringe plunger is pulled back to aspirate 20 mL of bone marrow. The cannula and syringe assembly are advanced 2 cm farther into the bone to allow for aspiration of a new location within the bone marrow cavity, and 20 mL of bone marrow are again aspirated. This is done a final time, advancing the trocar another 2 cm and aspirating a final 20 mL of bone marrow. The entire process should yield roughly 60 mL of bone marrow from one side. If desired, the same process can be repeated for the contralateral PSIS to yield a total of 120 mL of bone marrow from the 2 sites.

Based on our data, the average distance to the anterior ilium wall was 7 cm, but the shortest distance noted in this study was 5 cm. On the basis of the data presented, this technique allows for safe advancement based on even the shortest measured distance, without fear of puncturing the anterior ilium wall. Perforation could damage the femoral nerve and the internal or external iliac artery or vein that lie anterior to the ilium.

Continue to: We hypothesized that there...

 

DISCUSSION

We hypothesized that there would be an optimal angle of entry and maximal safe distance the trocar could advance through the ilium when aspirating. Because male and female pelvic anatomy differs, we also hypothesized that there would be differences in distance and size measurements for males and females. Our results supported our hypothesis that there is an ideal approach angle. The results also showed that the maximum distance the trocar can advance and the width of the PSIS table differ significantly between males and females.

Although pelvic anatomy differs between males and females, there should be an ideal entry angle that would allow maximum advancement into the ilium without perforating the anterior wall, which we defined as the approach angle. In our comparison of 26 MRI scans, we found that the approach angle did not differ significantly between the 2 groups (13 males, 13 females). This allows clinicians to enter the PSIS at roughly 24° medial to the parasagittal line, maximizing the space before puncturing into the anterior pelvis in either males or females.

If clinicians were to enter perpendicular to the patient’s PSIS, they would, on average, be able to advance only 3.20 cm before encountering the SI joint. When entering at 24° as we recommend, the average distance increases to 7.14 cm. Although the angle did not differ significantly, there was a significant difference between males and females in the length from the PSIS to the anterior wall, with males having 7.53 cm distance and females 6.74 cm. This is an important measurement because if the anterior ilium wall is punctured, the femoral nerve and the common, internal and external iliac arteries and veins could be damaged, resulting in retroperitoneal hemorrhage.

A fatality in 2001 in the United Kingdom led to a national audit of bone marrow aspiration and biopsies.^4-6 Although these procedures were done primarily for patients with cancer, hemorrhagic events were the most frequent and serious events. This audit led to the identification of many risk factors. Bain^4-6 conducted reviews of bone marrow aspirations and biopsies in the United Kingdom from 2002 to 2004. Of a total of 53,088 procedures conducted during that time frame, 48 (0.09%) adverse events occurred, with 29 (0.05%) being hemorrhagic events. Although infrequent, hemorrhagic adverse events represent significant morbidity. Reviews such as those conducted by Bain^4-6 highlight the importance of a study that helps determine the optimal parameters for aspiration to ensure safety and reliability.

Hernigou and colleagues^2,3 conducted studies analyzing different “sectors” in an attempt to develop a safe aspiration technique. They found that obese patients were at higher risk, and some sites of aspiration (sectors 1, 4, 5) had increased risk for perforation and damage to surrounding structures. Their sector 6, which incorporated the entirety of the PSIS table, was considered the safest, most reliable site for trocar introduction.^2,3 Hernigou and colleagues,² in comparing the bone mass of the sectors, also noted that sector 6 has the greatest bone thickness close to the entry point, making it the most favorable site. The PSIS is not just a point; it is more a “table.” The PSIS can be palpated posteriorly, but this is inaccurate and unreliable, particularly in larger individuals. The PSIS table can be identified on ultrasound before introducing the trocar, which is a more reliable method of landmark identification than palpation guidance, just as in ultrasound-guided injections^7-16 and procedures.^9,12,17-19

Continue to: If the PSIS is not accurately...

 

If the PSIS is not accurately identified, penetration laterally will result in entering the ilium wing, where it is quite narrow. We found the distance between the posterior ilium wall and the SI joint to be only 1.10 cm wide (Figure 3); we defined this area as the narrow corridor. Superior and lateral entry could damage the superior cluneal nerves coming over the iliac crest, which are located 6 cm lateral to the SI joint. Inferior and lateral entry 6 cm below the PSIS could reach the greater sciatic foramen, damaging the sacral plexus and superior gluteal artery and vein. If the entry slips above the PSIS over the pelvis, the trocar could enter the retroperitoneal space and damage the femoral nerve and common iliac artery and vein, leading to a retroperitoneal hemorrhage.^4-6,20

MSCs are found as perivascular cells and lie in the cortices of bones.²¹ Following the approach angle and directed line from the PSIS to the anterior ilium wall described in this study (Figures 1 and 2), the trocar would pass through the narrow corridor as it advances farther into the ilium. The minimum width of this corridor was measured in this study and, on average, was 1.10 cm wide from cortex to cortex (Figure 3). As the bone marrow is aspirated from this narrow corridor, the clinician is gathering MSCs from both the lateral and medial cortices of the ilium. By aspirating from a greater surface area of the cortices, it is believed that this will increase the total collection of MSCs.

CONCLUSION

Although there are reports in the literature that describe techniques for bone marrow aspiration from the iliac crest, the techniques are very general and vague regarding the ideal angles and methods. Studies have attempted to quantify the safest entry sites for aspiration but have not detailed ideal parameters for collection. Blind aspiration from the iliac crest can have serious implications if adverse events occur, and thus there is a need for a safe and reliable method of aspiration from the iliac crest. Ultrasound guidance to identify anatomy, as opposed to palpation guidance, ensures anatomic placement of the trocar while minimizing the risk of aspiration. Based on the measurements gathered in this study, an optimal angle of entry and safe distance of penetration have been identified. Using our data and relevant literature, we developed a technique for a safe, consistent, and reliable method of bone marrow aspiration out of the iliac crest.

ABSTRACT

Use of mesenchymal stem cells from bone marrow has gained significant popularity. The iliac crest has been determined to be an effective site for harvesting mesenchymal stem cells. Review of the literature reveals that multiple techniques are used to harvest bone marrow aspirate from the iliac crest, but the descriptions are based on the experience of various authors as opposed to studied anatomy. A safe, reliable, and reproducible method for aspiration has yet to be studied and described. We hypothesized that there would be an ideal angle and distance for aspiration that would be the safest, most consistent, and most reliable. Using magnetic resonance imaging (MRI), we reviewed 26 total lumbar spine MRI scans (13 males, 13 females) and found that an angle of 24° should be used when entering the most medial aspect of the posterior superior iliac spine (PSIS) and that this angle did not differ between the sexes. The distance that the trocar can advance after entry before hitting the anterior ilium wall varied significantly between males and females, being 7.53 cm in males and 6.74 cm in females. In addition, the size of the PSIS table was significantly different between males and females (1.20 cm and 0.96 cm, respectively). No other significant differences in the measurements gathered were found. Using the data gleaned from this study, we developed an aspiration technique. This method uses ultrasound to determine the location of the PSIS and the entry point on the PSIS. This contrasts with most techniques that use landmark palpation, which is known to be unreliable and inaccurate. The described technique for aspiration from the PSIS is safe, reliable, reproducible, and substantiated by data.

The iliac crest is an effective site for harvesting bone marrow stem cells. It allows for easy access and is superficial in most individuals, allowing for a relatively quick and simple procedure. Use of mesenchymal stem cells (MSCs) for treatment of orthopedic injuries has grown recently. Whereas overall use has increased, review of the literature reveals very few techniques for iliac crest aspiration,¹ but these are not based on anatomic relationships or studies. Hernigou and colleagues^2,3 attempted to quantitatively evaluate potential “sectors” allowing for safe aspiration using cadaver and computed tomographic reconstruction imaging. We used magnetic resonance imaging (MRI) to analyze aspiration parameters. Owing to the ilium’s anatomy, improper positioning or aspiration technique during aspiration can result in serious injury.^2,4-6 We hypothesized that there is an ideal angle and positioning for bone marrow aspiration from the posterior superior iliac spine (PSIS) that is safe, consistent, and reproducible. Although most aspiration techniques use landmark palpation, this is unreliable and inaccurate, especially when compared with ultrasound-guided injections^7-16 and procedures.^9,12,17-19 We describe our technique using ultrasound to visualize patient anatomy and accurately determine anatomic entry with the trocar.

METHODS

MRI scans of 26 patients (13 males, 13 females) were reviewed to determine average angles and distances. Axial T2-weighted views of the lumbar spine were used in all analyses. The sacroiliac (SI) joint angle was defined as the angle formed between the vector through the midline of the pelvis and the vector that is parallel to the SI joint. The approach angle was defined as the angle formed between the vector of the most medial aspect of the PSIS through the ilium to the anterior wall and the vector through the midline of the pelvis (Figure 1). 

The distance, in centimeters, of the PSIS to the anterior ilium wall was measured to determine the maximum distance the trocar can advance without puncturing the anterior ilium wall. The perpendicular distance from the PSIS table to the anterior aspect of the ilium was measured to determine the average depth the trocar could advance if the clinician entered it perpendicular to the patient’s coronal plane (Figure 2). 

The PSIS table width was defined as the width, in centimeters, of the PSIS posteriorly. The minimum width, in centimeters, of the ilium was measured at the thinnest region of the ilium to determine the degree of variation if the trocar were entered too far laterally and advanced to the anterior ilium wall (Figure 3). 

Means and standard deviations were calculated by sex and as a total. Student’s t test was used to determine statistical significance (P < .05).

Continue to: For the 13 males, the mean SI joint...

 

RESULTS

The results are reported in the Table. 

Table. Measurements of Patients Taken on Axial T2-Weighted Views of Lumbosacral MRI Scans^a

Patient	SI Joint Angle (°)	Approach Angle (°)	PSIS Table Width (cm)	PSIS to Anterior Ilium Wall (cm)	Perpendicular Distance PSIS to Anterior Joint (cm)	Post Ilium Wall to SI Joint Width (cm)
Males
1	28.80	19.50	1.24	8.80	4.16	1.52
2	31.80	27.60	1.70	7.89	3.49	1.02
3	33.70	27.70	1.12	8.14	3.15	1.28
4	23.70	26.40	0.95	6.66	3.22	0.65
5	35.90	28.40	0.84	7.60	2.57	0.95
6	33.80	29.30	1.20	7.73	2.34	0.90
7	30.30	21.20	1.36	8.44	3.95	1.18
8	34.50	20.40	1.53	7.08	3.98	1.56
9	28.70	24.00	1.34	8.19	3.51	1.31
10	22.40	20.10	1.37	7.30	3.87	1.28
11	33.60	20.80	0.88	6.43	3.26	0.94
12	48.50	31.00	1.15	6.69	2.97	1.38
13	20.20	20.90	0.94	6.95	3.79	1.05
Averages	31.22	24.41	1.20	7.53	3.40	1.16
Standard Deviation	7.18	4.11	0.26	0.75	0.56	0.26
Females
14	22.80	23.20	1.54	7.21	3.45	1.39
15	33.30	21.40	1.09	7.26	3.57	0.98
16	19.70	15.60	0.78	8.32	3.76	0.86
17	17.50	15.60	0.61	7.57	3.37	1.03
18	48.20	26.60	0.94	6.62	3.16	0.71
19	38.20	28.30	0.90	6.32	2.23	0.91
20	44.50	31.70	0.99	6.19	3.06	0.76
21	24.10	18.00	0.92	6.99	3.23	0.71
22	17.20	14.80	0.81	6.00	2.81	1.13
23	42.00	38.50	1.00	5.33	2.47	1.42
24	32.00	25.50	0.98	6.01	2.79	1.21
25	24.70	24.80	0.87	6.09	2.79	1.02
26	19.80	22.30	1.04	7.71	2.37	1.36
Averages	29.54	23.56	0.96	6.74	3.00	1.04
Standard Deviation	10.84	6.88	0.21	0.85	0.48	0.25
All patients Averages	30.38	23.98	1.08	7.14	3.20	1.10
Standard Deviation	9.05	5.57	0.26	0.88	0.55	0.26

^aStatistical significance is denoted as P < .02.

Abbreviations: MRI, magnetic resonance imaging; PSIS, posterior iliac spine; SI, sacroiliac.

For the 13 males, the mean SI joint angle was 31.22° ± 7.18° (range, 20.20° to 48.50°). The mean approach angle was 24.41° ± 4.11° (range, 19.50° to 31.00°). The mean PSIS table width was 1.20 cm ± 0.26 cm (range, 0.84 cm to 1.70 cm). The mean distance from the PSIS to the anterior ilium wall was 7.53 cm ± 0.75 cm (range, 6.43 cm to 8.80 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.40 cm ± 0.56 cm (range, 2.34 cm to 4.16 cm). The mean minimum width of the ilium to the SI joint was 1.16 cm ± 0.26 cm (range, 0.65 cm to 1.56 cm).

For the 13 females, the mean SI joint angle was 29.54° ± 10.84° (range, 17.20° to 48.20°). The mean approach angle was 23.56° ± 6.88° (range, 14.80° to 38.50°). The mean PSIS table width was 0.96 cm ± 0.21 cm (range, 0.61 cm to 1.54 cm). The mean distance from the PSIS to the anterior ilium wall was 6.74 cm ± 0.85 cm (range, 5.33 cm to 8.32 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.00 cm ± 0.48 cm (range, 2.23 cm to 3.76 cm). The mean minimum width of the ilium to the SI joint was 1.04 cm ± 0.25 cm (range, 0.71 cm to 1.42 cm).

For the 26 total patients, the mean SI joint angle was 30.38° ± 9.05° (range, 17.20° to 48.50°). The mean approach angle was 23.98° ± 5.57° (range, 14.80° to 38.50°). The mean PSIS table width was 1.08 cm ± 0.26 cm (range, 0.61 cm to 1.70 cm). The mean distance from the PSIS to the anterior ilium wall was 7.14 cm ± 0.88 cm (range, 5.33 cm to 8.80 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.20 cm ± 0.55 cm (range, 2.23 cm to 4.16 cm). The mean minimum width of the ilium to the SI joint was 1.10 cm ± 0.26 cm (range, 0.65 cm to 1.56 cm).

There was a statistically significant difference between the male and female groups for the maximum distance the trocar can be advanced from the PSIS to the anterior ilium wall (P < .02), and a statistically significant difference for the PSIS table width (P < .02). There were no significant differences between the male and female groups for the approach angle, the SI joint angle, the perpendicular distance from the PSIS to the anterior ilium, and the minimum width of the ilium to the SI joint.

Continue to: The patient is brought to the procedure...

 

TECHNIQUE: ILIAC CREST (PSIS) BONE MARROW ASPIRATION

The patient is brought to the procedure room and placed in a prone position. The donor site is prepared and draped in the usual sterile manner. Ultrasound is used to identify the median sacral crest in a short-axis view. The probe is then moved laterally to identify the PSIS (Figures 4A, 4B). 

The probe can be moved superiorly and inferiorly to determine the most prominent and central portion of the PSIS. The SI joint and ilium can also be visualized if needed.

The crosshairs on the ultrasound probe are used to mark the center lines of each plane. The central point marks the location of the PSIS. Alternatively, an in-plane technique can be used to place a spinal needle on the exact entry point on the PSIS. Once the PSIS and entry point are identified, the site is blocked with 10 mL of 0.5% ropivacaine.

Prior to introduction of the trocar, all instrumentation is primed with heparin and syringes are prepped with anticoagulant citrate dextrose solution, solution A. A stab incision is made at the site. The trocar is placed at the entry point, which should be centered in a superior-inferior plane and at the most medial point of the PSIS. Starting with the trocar vertical, the trocar is angled laterally 24° by dropping the hand medially toward the midline. No angulation cephalad or caudad is necessary, but cephalad must be avoided so as not to skive superiorly. This angle, which is recommended for both males and females, allows for the greatest distance the trocar can travel in bone before hitting the anterior ilium wall. A standard deviation of 5.57° is present, which should be considered. Steady pressure should be applied with a slight twisting motion on the PSIS. If advancement of the trocar is too difficult, a mallet or drill can be used to assist in penetration.

With the trocar advanced into the bone 1 cm, the trocar needle is removed while the cannula remains in place. The syringe is attached to the top of the cannula. The syringe plunger is pulled back to aspirate 20 mL of bone marrow. The cannula and syringe assembly are advanced 2 cm farther into the bone to allow for aspiration of a new location within the bone marrow cavity, and 20 mL of bone marrow are again aspirated. This is done a final time, advancing the trocar another 2 cm and aspirating a final 20 mL of bone marrow. The entire process should yield roughly 60 mL of bone marrow from one side. If desired, the same process can be repeated for the contralateral PSIS to yield a total of 120 mL of bone marrow from the 2 sites.

Based on our data, the average distance to the anterior ilium wall was 7 cm, but the shortest distance noted in this study was 5 cm. On the basis of the data presented, this technique allows for safe advancement based on even the shortest measured distance, without fear of puncturing the anterior ilium wall. Perforation could damage the femoral nerve and the internal or external iliac artery or vein that lie anterior to the ilium.

Continue to: We hypothesized that there...

 

DISCUSSION

We hypothesized that there would be an optimal angle of entry and maximal safe distance the trocar could advance through the ilium when aspirating. Because male and female pelvic anatomy differs, we also hypothesized that there would be differences in distance and size measurements for males and females. Our results supported our hypothesis that there is an ideal approach angle. The results also showed that the maximum distance the trocar can advance and the width of the PSIS table differ significantly between males and females.

Although pelvic anatomy differs between males and females, there should be an ideal entry angle that would allow maximum advancement into the ilium without perforating the anterior wall, which we defined as the approach angle. In our comparison of 26 MRI scans, we found that the approach angle did not differ significantly between the 2 groups (13 males, 13 females). This allows clinicians to enter the PSIS at roughly 24° medial to the parasagittal line, maximizing the space before puncturing into the anterior pelvis in either males or females.

If clinicians were to enter perpendicular to the patient’s PSIS, they would, on average, be able to advance only 3.20 cm before encountering the SI joint. When entering at 24° as we recommend, the average distance increases to 7.14 cm. Although the angle did not differ significantly, there was a significant difference between males and females in the length from the PSIS to the anterior wall, with males having 7.53 cm distance and females 6.74 cm. This is an important measurement because if the anterior ilium wall is punctured, the femoral nerve and the common, internal and external iliac arteries and veins could be damaged, resulting in retroperitoneal hemorrhage.

A fatality in 2001 in the United Kingdom led to a national audit of bone marrow aspiration and biopsies.^4-6 Although these procedures were done primarily for patients with cancer, hemorrhagic events were the most frequent and serious events. This audit led to the identification of many risk factors. Bain^4-6 conducted reviews of bone marrow aspirations and biopsies in the United Kingdom from 2002 to 2004. Of a total of 53,088 procedures conducted during that time frame, 48 (0.09%) adverse events occurred, with 29 (0.05%) being hemorrhagic events. Although infrequent, hemorrhagic adverse events represent significant morbidity. Reviews such as those conducted by Bain^4-6 highlight the importance of a study that helps determine the optimal parameters for aspiration to ensure safety and reliability.

Hernigou and colleagues^2,3 conducted studies analyzing different “sectors” in an attempt to develop a safe aspiration technique. They found that obese patients were at higher risk, and some sites of aspiration (sectors 1, 4, 5) had increased risk for perforation and damage to surrounding structures. Their sector 6, which incorporated the entirety of the PSIS table, was considered the safest, most reliable site for trocar introduction.^2,3 Hernigou and colleagues,² in comparing the bone mass of the sectors, also noted that sector 6 has the greatest bone thickness close to the entry point, making it the most favorable site. The PSIS is not just a point; it is more a “table.” The PSIS can be palpated posteriorly, but this is inaccurate and unreliable, particularly in larger individuals. The PSIS table can be identified on ultrasound before introducing the trocar, which is a more reliable method of landmark identification than palpation guidance, just as in ultrasound-guided injections^7-16 and procedures.^9,12,17-19

Continue to: If the PSIS is not accurately...

 

If the PSIS is not accurately identified, penetration laterally will result in entering the ilium wing, where it is quite narrow. We found the distance between the posterior ilium wall and the SI joint to be only 1.10 cm wide (Figure 3); we defined this area as the narrow corridor. Superior and lateral entry could damage the superior cluneal nerves coming over the iliac crest, which are located 6 cm lateral to the SI joint. Inferior and lateral entry 6 cm below the PSIS could reach the greater sciatic foramen, damaging the sacral plexus and superior gluteal artery and vein. If the entry slips above the PSIS over the pelvis, the trocar could enter the retroperitoneal space and damage the femoral nerve and common iliac artery and vein, leading to a retroperitoneal hemorrhage.^4-6,20

MSCs are found as perivascular cells and lie in the cortices of bones.²¹ Following the approach angle and directed line from the PSIS to the anterior ilium wall described in this study (Figures 1 and 2), the trocar would pass through the narrow corridor as it advances farther into the ilium. The minimum width of this corridor was measured in this study and, on average, was 1.10 cm wide from cortex to cortex (Figure 3). As the bone marrow is aspirated from this narrow corridor, the clinician is gathering MSCs from both the lateral and medial cortices of the ilium. By aspirating from a greater surface area of the cortices, it is believed that this will increase the total collection of MSCs.

CONCLUSION

Although there are reports in the literature that describe techniques for bone marrow aspiration from the iliac crest, the techniques are very general and vague regarding the ideal angles and methods. Studies have attempted to quantify the safest entry sites for aspiration but have not detailed ideal parameters for collection. Blind aspiration from the iliac crest can have serious implications if adverse events occur, and thus there is a need for a safe and reliable method of aspiration from the iliac crest. Ultrasound guidance to identify anatomy, as opposed to palpation guidance, ensures anatomic placement of the trocar while minimizing the risk of aspiration. Based on the measurements gathered in this study, an optimal angle of entry and safe distance of penetration have been identified. Using our data and relevant literature, we developed a technique for a safe, consistent, and reliable method of bone marrow aspiration out of the iliac crest.

ABSTRACT

Use of mesenchymal stem cells from bone marrow has gained significant popularity. The iliac crest has been determined to be an effective site for harvesting mesenchymal stem cells. Review of the literature reveals that multiple techniques are used to harvest bone marrow aspirate from the iliac crest, but the descriptions are based on the experience of various authors as opposed to studied anatomy. A safe, reliable, and reproducible method for aspiration has yet to be studied and described. We hypothesized that there would be an ideal angle and distance for aspiration that would be the safest, most consistent, and most reliable. Using magnetic resonance imaging (MRI), we reviewed 26 total lumbar spine MRI scans (13 males, 13 females) and found that an angle of 24° should be used when entering the most medial aspect of the posterior superior iliac spine (PSIS) and that this angle did not differ between the sexes. The distance that the trocar can advance after entry before hitting the anterior ilium wall varied significantly between males and females, being 7.53 cm in males and 6.74 cm in females. In addition, the size of the PSIS table was significantly different between males and females (1.20 cm and 0.96 cm, respectively). No other significant differences in the measurements gathered were found. Using the data gleaned from this study, we developed an aspiration technique. This method uses ultrasound to determine the location of the PSIS and the entry point on the PSIS. This contrasts with most techniques that use landmark palpation, which is known to be unreliable and inaccurate. The described technique for aspiration from the PSIS is safe, reliable, reproducible, and substantiated by data.

The iliac crest is an effective site for harvesting bone marrow stem cells. It allows for easy access and is superficial in most individuals, allowing for a relatively quick and simple procedure. Use of mesenchymal stem cells (MSCs) for treatment of orthopedic injuries has grown recently. Whereas overall use has increased, review of the literature reveals very few techniques for iliac crest aspiration,¹ but these are not based on anatomic relationships or studies. Hernigou and colleagues^2,3 attempted to quantitatively evaluate potential “sectors” allowing for safe aspiration using cadaver and computed tomographic reconstruction imaging. We used magnetic resonance imaging (MRI) to analyze aspiration parameters. Owing to the ilium’s anatomy, improper positioning or aspiration technique during aspiration can result in serious injury.^2,4-6 We hypothesized that there is an ideal angle and positioning for bone marrow aspiration from the posterior superior iliac spine (PSIS) that is safe, consistent, and reproducible. Although most aspiration techniques use landmark palpation, this is unreliable and inaccurate, especially when compared with ultrasound-guided injections^7-16 and procedures.^9,12,17-19 We describe our technique using ultrasound to visualize patient anatomy and accurately determine anatomic entry with the trocar.

METHODS

MRI scans of 26 patients (13 males, 13 females) were reviewed to determine average angles and distances. Axial T2-weighted views of the lumbar spine were used in all analyses. The sacroiliac (SI) joint angle was defined as the angle formed between the vector through the midline of the pelvis and the vector that is parallel to the SI joint. The approach angle was defined as the angle formed between the vector of the most medial aspect of the PSIS through the ilium to the anterior wall and the vector through the midline of the pelvis (Figure 1). 

The distance, in centimeters, of the PSIS to the anterior ilium wall was measured to determine the maximum distance the trocar can advance without puncturing the anterior ilium wall. The perpendicular distance from the PSIS table to the anterior aspect of the ilium was measured to determine the average depth the trocar could advance if the clinician entered it perpendicular to the patient’s coronal plane (Figure 2). 

The PSIS table width was defined as the width, in centimeters, of the PSIS posteriorly. The minimum width, in centimeters, of the ilium was measured at the thinnest region of the ilium to determine the degree of variation if the trocar were entered too far laterally and advanced to the anterior ilium wall (Figure 3). 

Means and standard deviations were calculated by sex and as a total. Student’s t test was used to determine statistical significance (P < .05).

Continue to: For the 13 males, the mean SI joint...

 

RESULTS

The results are reported in the Table. 

Table. Measurements of Patients Taken on Axial T2-Weighted Views of Lumbosacral MRI Scans^a

Patient	SI Joint Angle (°)	Approach Angle (°)	PSIS Table Width (cm)	PSIS to Anterior Ilium Wall (cm)	Perpendicular Distance PSIS to Anterior Joint (cm)	Post Ilium Wall to SI Joint Width (cm)
Males
1	28.80	19.50	1.24	8.80	4.16	1.52
2	31.80	27.60	1.70	7.89	3.49	1.02
3	33.70	27.70	1.12	8.14	3.15	1.28
4	23.70	26.40	0.95	6.66	3.22	0.65
5	35.90	28.40	0.84	7.60	2.57	0.95
6	33.80	29.30	1.20	7.73	2.34	0.90
7	30.30	21.20	1.36	8.44	3.95	1.18
8	34.50	20.40	1.53	7.08	3.98	1.56
9	28.70	24.00	1.34	8.19	3.51	1.31
10	22.40	20.10	1.37	7.30	3.87	1.28
11	33.60	20.80	0.88	6.43	3.26	0.94
12	48.50	31.00	1.15	6.69	2.97	1.38
13	20.20	20.90	0.94	6.95	3.79	1.05
Averages	31.22	24.41	1.20	7.53	3.40	1.16
Standard Deviation	7.18	4.11	0.26	0.75	0.56	0.26
Females
14	22.80	23.20	1.54	7.21	3.45	1.39
15	33.30	21.40	1.09	7.26	3.57	0.98
16	19.70	15.60	0.78	8.32	3.76	0.86
17	17.50	15.60	0.61	7.57	3.37	1.03
18	48.20	26.60	0.94	6.62	3.16	0.71
19	38.20	28.30	0.90	6.32	2.23	0.91
20	44.50	31.70	0.99	6.19	3.06	0.76
21	24.10	18.00	0.92	6.99	3.23	0.71
22	17.20	14.80	0.81	6.00	2.81	1.13
23	42.00	38.50	1.00	5.33	2.47	1.42
24	32.00	25.50	0.98	6.01	2.79	1.21
25	24.70	24.80	0.87	6.09	2.79	1.02
26	19.80	22.30	1.04	7.71	2.37	1.36
Averages	29.54	23.56	0.96	6.74	3.00	1.04
Standard Deviation	10.84	6.88	0.21	0.85	0.48	0.25
All patients Averages	30.38	23.98	1.08	7.14	3.20	1.10
Standard Deviation	9.05	5.57	0.26	0.88	0.55	0.26

^aStatistical significance is denoted as P < .02.

Abbreviations: MRI, magnetic resonance imaging; PSIS, posterior iliac spine; SI, sacroiliac.

For the 13 males, the mean SI joint angle was 31.22° ± 7.18° (range, 20.20° to 48.50°). The mean approach angle was 24.41° ± 4.11° (range, 19.50° to 31.00°). The mean PSIS table width was 1.20 cm ± 0.26 cm (range, 0.84 cm to 1.70 cm). The mean distance from the PSIS to the anterior ilium wall was 7.53 cm ± 0.75 cm (range, 6.43 cm to 8.80 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.40 cm ± 0.56 cm (range, 2.34 cm to 4.16 cm). The mean minimum width of the ilium to the SI joint was 1.16 cm ± 0.26 cm (range, 0.65 cm to 1.56 cm).

For the 13 females, the mean SI joint angle was 29.54° ± 10.84° (range, 17.20° to 48.20°). The mean approach angle was 23.56° ± 6.88° (range, 14.80° to 38.50°). The mean PSIS table width was 0.96 cm ± 0.21 cm (range, 0.61 cm to 1.54 cm). The mean distance from the PSIS to the anterior ilium wall was 6.74 cm ± 0.85 cm (range, 5.33 cm to 8.32 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.00 cm ± 0.48 cm (range, 2.23 cm to 3.76 cm). The mean minimum width of the ilium to the SI joint was 1.04 cm ± 0.25 cm (range, 0.71 cm to 1.42 cm).

For the 26 total patients, the mean SI joint angle was 30.38° ± 9.05° (range, 17.20° to 48.50°). The mean approach angle was 23.98° ± 5.57° (range, 14.80° to 38.50°). The mean PSIS table width was 1.08 cm ± 0.26 cm (range, 0.61 cm to 1.70 cm). The mean distance from the PSIS to the anterior ilium wall was 7.14 cm ± 0.88 cm (range, 5.33 cm to 8.80 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.20 cm ± 0.55 cm (range, 2.23 cm to 4.16 cm). The mean minimum width of the ilium to the SI joint was 1.10 cm ± 0.26 cm (range, 0.65 cm to 1.56 cm).

There was a statistically significant difference between the male and female groups for the maximum distance the trocar can be advanced from the PSIS to the anterior ilium wall (P < .02), and a statistically significant difference for the PSIS table width (P < .02). There were no significant differences between the male and female groups for the approach angle, the SI joint angle, the perpendicular distance from the PSIS to the anterior ilium, and the minimum width of the ilium to the SI joint.

Continue to: The patient is brought to the procedure...

 

TECHNIQUE: ILIAC CREST (PSIS) BONE MARROW ASPIRATION

The patient is brought to the procedure room and placed in a prone position. The donor site is prepared and draped in the usual sterile manner. Ultrasound is used to identify the median sacral crest in a short-axis view. The probe is then moved laterally to identify the PSIS (Figures 4A, 4B). 

The probe can be moved superiorly and inferiorly to determine the most prominent and central portion of the PSIS. The SI joint and ilium can also be visualized if needed.

The crosshairs on the ultrasound probe are used to mark the center lines of each plane. The central point marks the location of the PSIS. Alternatively, an in-plane technique can be used to place a spinal needle on the exact entry point on the PSIS. Once the PSIS and entry point are identified, the site is blocked with 10 mL of 0.5% ropivacaine.

Prior to introduction of the trocar, all instrumentation is primed with heparin and syringes are prepped with anticoagulant citrate dextrose solution, solution A. A stab incision is made at the site. The trocar is placed at the entry point, which should be centered in a superior-inferior plane and at the most medial point of the PSIS. Starting with the trocar vertical, the trocar is angled laterally 24° by dropping the hand medially toward the midline. No angulation cephalad or caudad is necessary, but cephalad must be avoided so as not to skive superiorly. This angle, which is recommended for both males and females, allows for the greatest distance the trocar can travel in bone before hitting the anterior ilium wall. A standard deviation of 5.57° is present, which should be considered. Steady pressure should be applied with a slight twisting motion on the PSIS. If advancement of the trocar is too difficult, a mallet or drill can be used to assist in penetration.

With the trocar advanced into the bone 1 cm, the trocar needle is removed while the cannula remains in place. The syringe is attached to the top of the cannula. The syringe plunger is pulled back to aspirate 20 mL of bone marrow. The cannula and syringe assembly are advanced 2 cm farther into the bone to allow for aspiration of a new location within the bone marrow cavity, and 20 mL of bone marrow are again aspirated. This is done a final time, advancing the trocar another 2 cm and aspirating a final 20 mL of bone marrow. The entire process should yield roughly 60 mL of bone marrow from one side. If desired, the same process can be repeated for the contralateral PSIS to yield a total of 120 mL of bone marrow from the 2 sites.

Based on our data, the average distance to the anterior ilium wall was 7 cm, but the shortest distance noted in this study was 5 cm. On the basis of the data presented, this technique allows for safe advancement based on even the shortest measured distance, without fear of puncturing the anterior ilium wall. Perforation could damage the femoral nerve and the internal or external iliac artery or vein that lie anterior to the ilium.

Continue to: We hypothesized that there...

 

DISCUSSION

We hypothesized that there would be an optimal angle of entry and maximal safe distance the trocar could advance through the ilium when aspirating. Because male and female pelvic anatomy differs, we also hypothesized that there would be differences in distance and size measurements for males and females. Our results supported our hypothesis that there is an ideal approach angle. The results also showed that the maximum distance the trocar can advance and the width of the PSIS table differ significantly between males and females.

Although pelvic anatomy differs between males and females, there should be an ideal entry angle that would allow maximum advancement into the ilium without perforating the anterior wall, which we defined as the approach angle. In our comparison of 26 MRI scans, we found that the approach angle did not differ significantly between the 2 groups (13 males, 13 females). This allows clinicians to enter the PSIS at roughly 24° medial to the parasagittal line, maximizing the space before puncturing into the anterior pelvis in either males or females.

If clinicians were to enter perpendicular to the patient’s PSIS, they would, on average, be able to advance only 3.20 cm before encountering the SI joint. When entering at 24° as we recommend, the average distance increases to 7.14 cm. Although the angle did not differ significantly, there was a significant difference between males and females in the length from the PSIS to the anterior wall, with males having 7.53 cm distance and females 6.74 cm. This is an important measurement because if the anterior ilium wall is punctured, the femoral nerve and the common, internal and external iliac arteries and veins could be damaged, resulting in retroperitoneal hemorrhage.

A fatality in 2001 in the United Kingdom led to a national audit of bone marrow aspiration and biopsies.^4-6 Although these procedures were done primarily for patients with cancer, hemorrhagic events were the most frequent and serious events. This audit led to the identification of many risk factors. Bain^4-6 conducted reviews of bone marrow aspirations and biopsies in the United Kingdom from 2002 to 2004. Of a total of 53,088 procedures conducted during that time frame, 48 (0.09%) adverse events occurred, with 29 (0.05%) being hemorrhagic events. Although infrequent, hemorrhagic adverse events represent significant morbidity. Reviews such as those conducted by Bain^4-6 highlight the importance of a study that helps determine the optimal parameters for aspiration to ensure safety and reliability.

Hernigou and colleagues^2,3 conducted studies analyzing different “sectors” in an attempt to develop a safe aspiration technique. They found that obese patients were at higher risk, and some sites of aspiration (sectors 1, 4, 5) had increased risk for perforation and damage to surrounding structures. Their sector 6, which incorporated the entirety of the PSIS table, was considered the safest, most reliable site for trocar introduction.^2,3 Hernigou and colleagues,² in comparing the bone mass of the sectors, also noted that sector 6 has the greatest bone thickness close to the entry point, making it the most favorable site. The PSIS is not just a point; it is more a “table.” The PSIS can be palpated posteriorly, but this is inaccurate and unreliable, particularly in larger individuals. The PSIS table can be identified on ultrasound before introducing the trocar, which is a more reliable method of landmark identification than palpation guidance, just as in ultrasound-guided injections^7-16 and procedures.^9,12,17-19

Continue to: If the PSIS is not accurately...

 

If the PSIS is not accurately identified, penetration laterally will result in entering the ilium wing, where it is quite narrow. We found the distance between the posterior ilium wall and the SI joint to be only 1.10 cm wide (Figure 3); we defined this area as the narrow corridor. Superior and lateral entry could damage the superior cluneal nerves coming over the iliac crest, which are located 6 cm lateral to the SI joint. Inferior and lateral entry 6 cm below the PSIS could reach the greater sciatic foramen, damaging the sacral plexus and superior gluteal artery and vein. If the entry slips above the PSIS over the pelvis, the trocar could enter the retroperitoneal space and damage the femoral nerve and common iliac artery and vein, leading to a retroperitoneal hemorrhage.^4-6,20

MSCs are found as perivascular cells and lie in the cortices of bones.²¹ Following the approach angle and directed line from the PSIS to the anterior ilium wall described in this study (Figures 1 and 2), the trocar would pass through the narrow corridor as it advances farther into the ilium. The minimum width of this corridor was measured in this study and, on average, was 1.10 cm wide from cortex to cortex (Figure 3). As the bone marrow is aspirated from this narrow corridor, the clinician is gathering MSCs from both the lateral and medial cortices of the ilium. By aspirating from a greater surface area of the cortices, it is believed that this will increase the total collection of MSCs.

CONCLUSION

Although there are reports in the literature that describe techniques for bone marrow aspiration from the iliac crest, the techniques are very general and vague regarding the ideal angles and methods. Studies have attempted to quantify the safest entry sites for aspiration but have not detailed ideal parameters for collection. Blind aspiration from the iliac crest can have serious implications if adverse events occur, and thus there is a need for a safe and reliable method of aspiration from the iliac crest. Ultrasound guidance to identify anatomy, as opposed to palpation guidance, ensures anatomic placement of the trocar while minimizing the risk of aspiration. Based on the measurements gathered in this study, an optimal angle of entry and safe distance of penetration have been identified. Using our data and relevant literature, we developed a technique for a safe, consistent, and reliable method of bone marrow aspiration out of the iliac crest.

The majority of hospitalized patients worldwide have at least one peripheral intravenous catheter (PIVC),¹ making PIVC insertion one of the most common clinical procedures. In the United States, physicians, advanced practitioners, and nurses insert over 300 million of these devices in hospitalized patients annually.² Despite their prevalence, PIVCs are associated with high rates of complications, including insertion difficulty, phlebitis, infiltration, occlusion, dislodgment, and catheter-associated bloodstream infection (CABSI), known to increase morbidity and mortality risk.^2-9 Up to 90% of PIVCs are prematurely removed owing to failure before planned replacement or before intravenous (IV) therapy completion.^3-6,10-12

PIVC complication and failure commonly triggers insertion of a replacement device and can entail significant costs.^2-4 One example is PIVC-related CABSI, where treatment costs have been estimated to be between US$35,000 and US$56,000 per patient.^6,13 Another important consideration is the pain and anxiety experienced by patients who need a replacement device, particularly those with difficult vascular access, who may require multiple cannulation attempts to replace a PIVC.^12,14-16 In developing nations, serious adverse events related to PIVCs are even more concerning, because hospital acquired infection rates and associated mortality are nearly 20 times greater than in developed nations.¹⁷

A number of evidence-based interventions have been suggested to reduce PIVC failure rates. In addition to optimal hand hygiene when inserting or accessing a PIVC to prevent infection,¹⁸ recommended interventions include placement of the PIVC in an area of non-flexion such as the forearm to provide stability for the device and to reduce patient discomfort, securing the PIVC to reduce movement of the catheter at the insertion site and within the blood vessel, and use of occlusive dressings that reduce the risk of external contamination of the PIVC site.^11,19,20 Best practice guidelines also recommend the prompt removal of devices that are symptomatic (when phlebitis or other complications are suspected) and when the catheter is no longer required.^21,22

Recent evidence has demonstrated that catheter size can have an impact on device survival rates. In adults, large-bore catheters of 18 gauge (G) or higher were found to have an increased rate of thrombosis, and smaller-bore catheters of 22G or lower (in adults) were found to have higher rates of dislodgment and occlusion/infiltration. The catheter size recommended for adults based on the latest evidence for most clinical applications is 20G.^3,20,23,24 In addition, the documentation of insertion, maintenance, and removal of PIVCs in the medical record is a requirement in most healthcare facilities worldwide and is recommended by best practice guidelines; however, adherence remains a challenge.^1,19

The concerning prevalence of PIVC-related complications and the lack of comparative data internationally on organizational compliance with best practice guidelines formed the rationale for this study. Our study aim was to describe the insertion characteristics, management practices, and outcomes of PIVCs internationally and to compare these variables to recommended best practice.

MATERIALS AND METHODS

Study Design and Participants

In this international cross-sectional study, we recruited hospitals through professional networks, including vascular access, infection prevention, safety and quality, nursing, and hospital associations (Appendix 2). Healthcare organizations, government health departments, and intravascular device suppliers were informed of the study and requested to further disseminate information through their networks. A study website was developed,²⁵ and social media outlets, including Twitter^®, LinkedIn^®, and Facebook^®, were used to promote the study.

 

Approval was granted by the Griffith University Human Research Ethics Committee in Australia (reference number NRS/34/13/HREC). In addition, evidence of study site and local institutional review board/ethics committee approval was required prior to study commencement. Each participating site agreed to follow the study protocol and signed an authorship agreement form. No financial support was provided to any site.

Hospitalized adult and pediatric patients with a PIVC in situ on the day of the study were eligible for inclusion. Sample size was determined by local capacity. Hospitals were encouraged to audit their entire institution if possible; however, data were accepted from as little as one ward. Data collectors comprised nurses and doctors with experience in PIVC assessment. They were briefed on the study protocol and data collection forms by the local site coordinator, and they were supported by an overall global coordinator. Clinicians assessed the PIVC insertion site and accessed hospital records to collect data related to PIVC insertion, concurrent medications, and IV fluid orders. Further clarification of data was obtained if necessary by the clinicians from the patients and treating staff. No identifiable patient information was collected.

Data Collection

To assess whether clinical facilities were following best practice recommendations, the study team developed three data collection forms to collect information regarding site characteristics (site questionnaire), track participant recruitment (screening log), and collect data regarding PIVC characteristics and management practices (case report form [CRF]). All forms were internally and externally validated following a pilot study involving 14 sites in 13 countries.¹

The CRF included variables used to assess best practice interventions, such as catheter insertion characteristics (date and time, reason, location, profession of inserter, anatomical site of placement), catheter type (gauge, brand, and product), insertion site assessment (adverse symptoms, dressing type and integrity), and information related to the IV therapy (types of IV fluids and medications, flushing solutions). Idle PIVCs were defined as not being used for blood sampling or IV therapy in the preceding 24 h.

Data collection forms were translated into 15 languages by professional translators and back-translated for validity. Translation of some languages included additional rigor. For example, Spanish-speaking members from the Spanish mainland as well as from South America were employed so that appropriate synonyms were used to capture local terms and practice. Three options were provided for data entry: directly into a purpose-developed electronic database (Lime Survey^® Project, Hamburg, Germany); on paper, then transcribed into the survey database at a later time by the hospital site; or paper entry then sent (via email or post) to the coordinating center for data entry. Once cleaned and collated, all data were provided to each participating hospital to confirm accuracy and for site use in local quality improvement processes. Data were collected between June 1, 2014 and July 31, 2015.

Statistical Analysis

All data management was undertaken using SAS statistical software (SAS Institute Inc., Cary NC, USA). Results are presented for eight geographical regions using descriptive statistics (frequencies, percentages, and 95% CIs) for the variables of interest. To assess trends in catheter dwell time and rates of phlebitis, Poisson regression was used. All analyses were undertaken using the R language for statistical analysis (R Core Team, Vienna, Austria). The (STROBE (Strengthening the Reporting of Observational Studies in Epidemiology statement) guidelines for cross-sectional studies were followed, and results are presented according to these recommendations.²⁶

RESULTS

Of the 415 hospitals that participated in this study, 406 had patients with PIVCs on the day of the study (the others being small rural centers). Thus, a total of 40,620 PIVCs in 38,161 patients from 406 hospitals in 51 countries were assessed, with no more than 5% missing data for any CRF question. There were 2459 patients (6.1%) with two or more PIVCs concurrently in situ. The median patient age was 59 y (interquartile range [IQR], 37–74 y), and just over half were male (n = 20,550, 51%). Hospital size ranged from fewer than 10 beds to over 1,000 beds, and hospitals were located in rural, regional, and metropolitan districts. The majority of countries (n = 31, 61%) contributed multiple sites, the highest being Australia with 79 hospitals. Countries with the most PIVCs studied were Spain (n = 5,553, 14%) and the United States (n = 5,048, 12%).

General surgical (n = 15,616, 39%) and medical (n = 15,448, 38%) patients represented most of the population observed. PIVCs were inserted primarily in general wards or clinics (n = 22,167, 55%) or in emergency departments (n = 7,388, 18%; Table) and for the administration of IV medication (n = 28,571, 70%) and IV fluids (n = 7,093, 18%; Table).

Globally, nurses were the primary PIVC inserters (n = 28,575, 71%); however, Australia/New Zealand had only 26% (n = 1,518) of PIVCs inserted by this group (Table). Only about one-third of PIVCs were placed in an area of non-flexion (forearm, n = 12,675, 31%, Table) the majority (n = 27,856, 69%) were placed in non-recommended anatomical sites (Figure 1). Most PIVCs were placed in the hand (n = 13,265, 32.7%) followed by the antecubital veins (n = 6176, 15.2%) and the wrist (n = 5,465, 13.5%). Site selection varied widely across the regions; 29% (n = 1686) of PIVCs in Australia/New Zealand were inserted into the antecubital veins, twice the study group average. Over half of the PIVCs inserted in the Middle East were placed in the hand (n = 295, 56%). This region also had the highest prevalence of devices placed in nonrecommended sites (n = 416, 79%; Figure 1).

The majority of PIVCs (n = 27,192, 67%; Table) were of recommended size (20–22G); however, some devices were observed to be large (14–18G; n = 6,802, 17%) or small (24-26g; n = 4,869, 12%) in adults. In Asia, 41% (n = 2,617) of devices inserted were 24-26G, more than three times the global rate. Half of all devices in Asia (n = 3,077, 48%) and the South Pacific (n = 67, 52%) were of a size not recommended for routine IV therapy (Figure 2).

The primary dressing material used was a transparent dressing (n = 31,596, 77.8%; Table); however, nearly 1 in 5 dressings used had either nonsterile tape alone (n = 5,169, 13%; Appendix 4), or a sterile gauze and tape (n = 2,592, 6%; Appendix 4.1). We found a wide variation in the use of nonsterile tape, including 1 in every 3 devices in South America dressed with nonsterile tape (n = 714, 30%) and a larger proportion in Africa (n = 543, 19%) and Europe (n = 3,056, 18%). Nonsterile tape was rarely used in North America and Australia/New Zealand. Although most PIVC dressings were clean, dry, and intact (n = 31,786, 79%; Table), one-fifth overall were compromised (moist, soiled, and/or lifting off the skin). Compromised dressings (Appendix 4.2) were more prevalent in Australia/New Zealand (n = 1,448; 25%) and in Africa (n = 707, 25%) than elsewhere.

Ten percent of PIVCs (n = 4,204) had signs and/or symptoms suggestive of phlebitis (characterized by pain, redness and/or swelling at the insertion site; Appendix 4.3). The highest prevalence of phlebitis occurred in Asia (n = 1,021, 16%), Africa (n = 360, 13%), and South America (n = 284, 12%). Pain and/or redness were the most common phlebitis symptoms. We found no association between dwell time of PIVCs and phlebitis rates (P = .085). Phlebitis rates were 12% (Days 1-3; n = 15,625), 16% (Days 4-7; n = 3,348), 10% (Days 8-21; n = 457), and 13% (Day21+; n = 174). Nearly 10% (n = 3,879) of catheters were observed to have signs of malfunction such as blood in the infusion tubing, leaking at the insertion site, or dislodgment (Appendix 4.4).

We observed 14% (n = 5,796) of PIVCs to be idle (Appendix 4.5), defined as not used in the preceding 24 h. Nearly one-fourth of all devices in North America (n = 1,230, 23%) and Australia/New Zealand (n = 1,335, 23%) were idle. PIVC documentation in hospital records was also poor, nearly half of all PIVCs (n = 19,768, 49%) had no documented date and time of insertion. The poorest compliance was in Australia/New Zealand (n = 3,428, 59%; Appendix 4.6). We also observed that 1 in 10 PIVCs had no documentation regarding who inserted the PIVC (n = 3,905). Thirty-six percent of PIVCs (n = 14,787) had no documented assessment of the PIVC site on the day of review (Appendix 4.7), including over half of all PIVCs in Asia (n = 3,364, 52%). Overall, the median dwell at the time of assessment for PIVCs with insertion date/time documented was 1.5 d (IQR, 1.0–2.5 d).

 

DISCUSSION

This international assessment of more than 40,000 PIVCs in 51 countries provides great insight into device characteristics and variation in management practices. Predominantly, PIVCs were inserted by nurses in the general ward environment for IV medication. One in ten PIVCs had at least one symptom of phlebitis, one in ten were dysfunctional, one in five PIVC dressings were compromised, and one in six PIVCs had not been used in the preceding 24 h. Nearly half of the PIVCs audited had the insertion date and time missing.

Regional variation was found in the professions inserting PIVCs, as well as in anatomical placement. In Australia/New Zealand, the proportion of nurses inserting PIVCs was much lower than the study group average (26% vs 71%). Because these countries contributed a substantial number of hospitals to the study, this seems a representative finding and suggests a need for education targeted at nurses for PIVC insertion in this region. The veins in the forearm are recommended as optimal for PIVC insertion in adults, rather than areas of high flexion, because the forearm provides a wide surface area to secure and dress PIVCs. Forearm placement can reduce pain during catheter dwell as well as decrease the risk of accidental removal or occlusion.^3,19,27 We found only one-third of PIVCs were placed in the forearm, with most placed in the hand, antecubital veins, or wrist. This highlights an inconsistency with published recommendations and suggests that additional training and technology are required so that staff can better identify and insert PIVCs in the forearm for other than very short-term (procedural) PIVCp;s.¹⁹

Phlebitis triggering PIVC failure remains a global clinical challenge with numerous phlebitis definitions and varied assessment techniques.¹⁰ The prevalence of phlebitis has been difficult to approximate with varying estimates and definitions in the literature; however, it remains a key predictor of PIVC failure.^6,10 Identification of this complication and prompt removal of the device is critical for patient comfort and reducing CABSI risk.^5,28 The overall prevalence of phlebitis signs or symptoms (defined in this study as having one or more signs of redness, swelling, or pain surrounding the insertion site) was just over 10%, with pain and/or redness being most prevalent. These compromised PIVCs had not been removed as is recommended for such complications.^19,28 Considering that our study was a snapshot at only one time point, the per-catheter incidence of phlebitis would be even higher; interestingly, among PIVCs with a documented insertion date and time, we observed that dwell time did not influence phlebitis rates.

Another concern is that nearly 10% (n = 3,879) of PIVCs were malfunctioning (eg, leaking) but were still in place. To bring these problems into context, around 2 billion PIVCs are used annually worldwide; as a consequence, millions of patients suffer from painful or malfunctioning PIVCs staff had not responded.^1,29 The placement of large-bore catheters, and smaller-gauge ones in adults, is known to increase the incidence of malfunction that leads to failure. There are a number of sound clinical reasons for the use of large-bore (eg, resuscitation and rapid fluid replacement) or small-bore (eg, difficult venous access with small superficial veins only visible and palpable) catheters. However, it would be expected that only a small proportion of patients would require these devices, and not one in three devices as we identified. This finding suggests that some PIVCs were inappropriate in size for general IV therapy and may reflect antiquated hospital policies for some clinical cohorts.^30,31

Overall, transparent dressings were used to cover the PIVC, but a number of patients were observed to have a sterile gauze and tape dressing (n = 2,592, 6%). Although the latter is less common, both dressing approaches are recommended in clinical practice guidelines because there is a lack of high-quality evidence regarding which is superior.^21,22,32 Of concern was the use of nonsterile tape to dress the PIVC (n = 5,169, 12.7%). We found the prevalence of nonsterile tape use to be higher in lower-resourced countries in South America (n = 714, 30%), Africa (n = 543, 19%) and Europe (n = 3,056, 18%) and this was likely related to institutional cost reduction practices.

This finding illustrates an important issue regarding proper PIVC care and management practices in developing nations. It is widely known that access to safe health care in lower-resourced nations is challenging and that rates of mortality related to healthcare-associated infections are much higher. Thus, the differences we found in PIVC management practices in these countries are not surprising.^33,34 International health networks such as the Infection Control Africa Network, the International Federation of Infection Control, and the Centers for Disease Control and Prevention can have great influence on ministries of health and clinicians in these countries to develop coordinated efforts for safe and sustainable IV practices to reduce the burden of hospital-acquired infections and related morbidity and mortality.

We found that 14% of all PIVCs had no documented IV medication or IV fluid administered in the previous 24 h, strongly indicating that they were no longer needed. Australia/New Zealand, Europe, and North America were observed to have a higher prevalence of idle catheters than the remaining regions. This suggests that an opportunity exists to develop surveillance systems that better identify idle devices for prompt removal to reduce infection risk and patient discomfort. Several randomized controlled trials, a Cochrane review, and clinical practice guidelines recommend prompt removal of PIVCs when not required, if there are any complications, or if the PIVC was inserted urgently without an aseptic insertion technique.^21,28,35,36 Idle PIVCs have been implicated in adverse patient outcomes, including phlebitis and CABSI.^13,27

The substantial proportion of patients with a PIVC in this study who had no clinical indication for a PIVC, a symptomatic insertion site, malfunctioning catheter, and suboptimal dressing quality suggests the need for physicians, advanced practitioners, and nurses to adopt evidence-based PIVC insertion and maintenance bundles and supporting checklists to reduce the prevalence of PIVC complications.^19,21,38-40 Recommended strategies for inclusion in PIVC maintenance bundles are prompt removal of symptomatic and/or idle catheters, hand hygiene prior to accessing the catheter, regular assessment of the device, and replacement of suboptimal dressings.^41,42 This approach should be implemented across all clinical specialties involved in PIVC insertion and care.

Our study findings need to be considered within the context of some limitations. The cross-sectional design prevented follow-up of PIVCs until removal to collect outcomes, including subsequent PIVC complications and/or failure, following the study observation. Ideally, data collection could have included patient-level preferences for PIVC insertion, history of PIVC use and/or failure, the number of PIVC insertion attempts, and the number of PIVCs used during that hospitalization. However, a cohort study of this magnitude was not feasible, particularly because all sites contributed staff time to complete the data collection. Only half of all initially registered sites eventually participated in the study; reasons for not participating were cited as local workload constraints and/or difficulties in applying for local approvals. Although efforts to enroll hospitals worldwide were exhaustive, our sample was not randomly selected but relied on self-selection and so is not representative, particularly for countries that contributed only one hospital site. Caution is also required when comparing inter regional differences, particularly developing regions, because better-resourced/academic sites were possibly over represented in the sample. Nevertheless, PIVC variables differed significantly between participating hospitals, suggesting that the data represent a reasonable reflection of hospital variability.

 

CONCLUSIONS

On the basis of this international investigation, we report variations in the characteristics, management practices, and outcomes of PIVCs inserted in hospital patients from 51 countries. Many PIVCs were idle, symptomatic, had substandard dressings, and were inserted in suboptimal anatomical sites. Despite international best practice guidelines, a large number of patients had PIVCs that were already failing or at risk of complications, including infection. A stronger focus is needed on compliance with PIVC insertion and management guidelines; better surveillance of PIVC sites; and improved assessment, decision-making, and documentation.

Acknowledgements

We are extremely grateful to colleagues from across the globe who committed their time and effort to this study (for full details of countries and team members see Appendix 1).

The majority of hospitalized patients worldwide have at least one peripheral intravenous catheter (PIVC),¹ making PIVC insertion one of the most common clinical procedures. In the United States, physicians, advanced practitioners, and nurses insert over 300 million of these devices in hospitalized patients annually.² Despite their prevalence, PIVCs are associated with high rates of complications, including insertion difficulty, phlebitis, infiltration, occlusion, dislodgment, and catheter-associated bloodstream infection (CABSI), known to increase morbidity and mortality risk.^2-9 Up to 90% of PIVCs are prematurely removed owing to failure before planned replacement or before intravenous (IV) therapy completion.^3-6,10-12

PIVC complication and failure commonly triggers insertion of a replacement device and can entail significant costs.^2-4 One example is PIVC-related CABSI, where treatment costs have been estimated to be between US$35,000 and US$56,000 per patient.^6,13 Another important consideration is the pain and anxiety experienced by patients who need a replacement device, particularly those with difficult vascular access, who may require multiple cannulation attempts to replace a PIVC.^12,14-16 In developing nations, serious adverse events related to PIVCs are even more concerning, because hospital acquired infection rates and associated mortality are nearly 20 times greater than in developed nations.¹⁷

A number of evidence-based interventions have been suggested to reduce PIVC failure rates. In addition to optimal hand hygiene when inserting or accessing a PIVC to prevent infection,¹⁸ recommended interventions include placement of the PIVC in an area of non-flexion such as the forearm to provide stability for the device and to reduce patient discomfort, securing the PIVC to reduce movement of the catheter at the insertion site and within the blood vessel, and use of occlusive dressings that reduce the risk of external contamination of the PIVC site.^11,19,20 Best practice guidelines also recommend the prompt removal of devices that are symptomatic (when phlebitis or other complications are suspected) and when the catheter is no longer required.^21,22

Recent evidence has demonstrated that catheter size can have an impact on device survival rates. In adults, large-bore catheters of 18 gauge (G) or higher were found to have an increased rate of thrombosis, and smaller-bore catheters of 22G or lower (in adults) were found to have higher rates of dislodgment and occlusion/infiltration. The catheter size recommended for adults based on the latest evidence for most clinical applications is 20G.^3,20,23,24 In addition, the documentation of insertion, maintenance, and removal of PIVCs in the medical record is a requirement in most healthcare facilities worldwide and is recommended by best practice guidelines; however, adherence remains a challenge.^1,19

The concerning prevalence of PIVC-related complications and the lack of comparative data internationally on organizational compliance with best practice guidelines formed the rationale for this study. Our study aim was to describe the insertion characteristics, management practices, and outcomes of PIVCs internationally and to compare these variables to recommended best practice.

MATERIALS AND METHODS

Study Design and Participants

In this international cross-sectional study, we recruited hospitals through professional networks, including vascular access, infection prevention, safety and quality, nursing, and hospital associations (Appendix 2). Healthcare organizations, government health departments, and intravascular device suppliers were informed of the study and requested to further disseminate information through their networks. A study website was developed,²⁵ and social media outlets, including Twitter^®, LinkedIn^®, and Facebook^®, were used to promote the study.

 

Approval was granted by the Griffith University Human Research Ethics Committee in Australia (reference number NRS/34/13/HREC). In addition, evidence of study site and local institutional review board/ethics committee approval was required prior to study commencement. Each participating site agreed to follow the study protocol and signed an authorship agreement form. No financial support was provided to any site.

Hospitalized adult and pediatric patients with a PIVC in situ on the day of the study were eligible for inclusion. Sample size was determined by local capacity. Hospitals were encouraged to audit their entire institution if possible; however, data were accepted from as little as one ward. Data collectors comprised nurses and doctors with experience in PIVC assessment. They were briefed on the study protocol and data collection forms by the local site coordinator, and they were supported by an overall global coordinator. Clinicians assessed the PIVC insertion site and accessed hospital records to collect data related to PIVC insertion, concurrent medications, and IV fluid orders. Further clarification of data was obtained if necessary by the clinicians from the patients and treating staff. No identifiable patient information was collected.

Data Collection

To assess whether clinical facilities were following best practice recommendations, the study team developed three data collection forms to collect information regarding site characteristics (site questionnaire), track participant recruitment (screening log), and collect data regarding PIVC characteristics and management practices (case report form [CRF]). All forms were internally and externally validated following a pilot study involving 14 sites in 13 countries.¹

The CRF included variables used to assess best practice interventions, such as catheter insertion characteristics (date and time, reason, location, profession of inserter, anatomical site of placement), catheter type (gauge, brand, and product), insertion site assessment (adverse symptoms, dressing type and integrity), and information related to the IV therapy (types of IV fluids and medications, flushing solutions). Idle PIVCs were defined as not being used for blood sampling or IV therapy in the preceding 24 h.

Data collection forms were translated into 15 languages by professional translators and back-translated for validity. Translation of some languages included additional rigor. For example, Spanish-speaking members from the Spanish mainland as well as from South America were employed so that appropriate synonyms were used to capture local terms and practice. Three options were provided for data entry: directly into a purpose-developed electronic database (Lime Survey^® Project, Hamburg, Germany); on paper, then transcribed into the survey database at a later time by the hospital site; or paper entry then sent (via email or post) to the coordinating center for data entry. Once cleaned and collated, all data were provided to each participating hospital to confirm accuracy and for site use in local quality improvement processes. Data were collected between June 1, 2014 and July 31, 2015.

Statistical Analysis

All data management was undertaken using SAS statistical software (SAS Institute Inc., Cary NC, USA). Results are presented for eight geographical regions using descriptive statistics (frequencies, percentages, and 95% CIs) for the variables of interest. To assess trends in catheter dwell time and rates of phlebitis, Poisson regression was used. All analyses were undertaken using the R language for statistical analysis (R Core Team, Vienna, Austria). The (STROBE (Strengthening the Reporting of Observational Studies in Epidemiology statement) guidelines for cross-sectional studies were followed, and results are presented according to these recommendations.²⁶

RESULTS

Of the 415 hospitals that participated in this study, 406 had patients with PIVCs on the day of the study (the others being small rural centers). Thus, a total of 40,620 PIVCs in 38,161 patients from 406 hospitals in 51 countries were assessed, with no more than 5% missing data for any CRF question. There were 2459 patients (6.1%) with two or more PIVCs concurrently in situ. The median patient age was 59 y (interquartile range [IQR], 37–74 y), and just over half were male (n = 20,550, 51%). Hospital size ranged from fewer than 10 beds to over 1,000 beds, and hospitals were located in rural, regional, and metropolitan districts. The majority of countries (n = 31, 61%) contributed multiple sites, the highest being Australia with 79 hospitals. Countries with the most PIVCs studied were Spain (n = 5,553, 14%) and the United States (n = 5,048, 12%).

General surgical (n = 15,616, 39%) and medical (n = 15,448, 38%) patients represented most of the population observed. PIVCs were inserted primarily in general wards or clinics (n = 22,167, 55%) or in emergency departments (n = 7,388, 18%; Table) and for the administration of IV medication (n = 28,571, 70%) and IV fluids (n = 7,093, 18%; Table).

Globally, nurses were the primary PIVC inserters (n = 28,575, 71%); however, Australia/New Zealand had only 26% (n = 1,518) of PIVCs inserted by this group (Table). Only about one-third of PIVCs were placed in an area of non-flexion (forearm, n = 12,675, 31%, Table) the majority (n = 27,856, 69%) were placed in non-recommended anatomical sites (Figure 1). Most PIVCs were placed in the hand (n = 13,265, 32.7%) followed by the antecubital veins (n = 6176, 15.2%) and the wrist (n = 5,465, 13.5%). Site selection varied widely across the regions; 29% (n = 1686) of PIVCs in Australia/New Zealand were inserted into the antecubital veins, twice the study group average. Over half of the PIVCs inserted in the Middle East were placed in the hand (n = 295, 56%). This region also had the highest prevalence of devices placed in nonrecommended sites (n = 416, 79%; Figure 1).

The majority of PIVCs (n = 27,192, 67%; Table) were of recommended size (20–22G); however, some devices were observed to be large (14–18G; n = 6,802, 17%) or small (24-26g; n = 4,869, 12%) in adults. In Asia, 41% (n = 2,617) of devices inserted were 24-26G, more than three times the global rate. Half of all devices in Asia (n = 3,077, 48%) and the South Pacific (n = 67, 52%) were of a size not recommended for routine IV therapy (Figure 2).

The primary dressing material used was a transparent dressing (n = 31,596, 77.8%; Table); however, nearly 1 in 5 dressings used had either nonsterile tape alone (n = 5,169, 13%; Appendix 4), or a sterile gauze and tape (n = 2,592, 6%; Appendix 4.1). We found a wide variation in the use of nonsterile tape, including 1 in every 3 devices in South America dressed with nonsterile tape (n = 714, 30%) and a larger proportion in Africa (n = 543, 19%) and Europe (n = 3,056, 18%). Nonsterile tape was rarely used in North America and Australia/New Zealand. Although most PIVC dressings were clean, dry, and intact (n = 31,786, 79%; Table), one-fifth overall were compromised (moist, soiled, and/or lifting off the skin). Compromised dressings (Appendix 4.2) were more prevalent in Australia/New Zealand (n = 1,448; 25%) and in Africa (n = 707, 25%) than elsewhere.

Ten percent of PIVCs (n = 4,204) had signs and/or symptoms suggestive of phlebitis (characterized by pain, redness and/or swelling at the insertion site; Appendix 4.3). The highest prevalence of phlebitis occurred in Asia (n = 1,021, 16%), Africa (n = 360, 13%), and South America (n = 284, 12%). Pain and/or redness were the most common phlebitis symptoms. We found no association between dwell time of PIVCs and phlebitis rates (P = .085). Phlebitis rates were 12% (Days 1-3; n = 15,625), 16% (Days 4-7; n = 3,348), 10% (Days 8-21; n = 457), and 13% (Day21+; n = 174). Nearly 10% (n = 3,879) of catheters were observed to have signs of malfunction such as blood in the infusion tubing, leaking at the insertion site, or dislodgment (Appendix 4.4).

We observed 14% (n = 5,796) of PIVCs to be idle (Appendix 4.5), defined as not used in the preceding 24 h. Nearly one-fourth of all devices in North America (n = 1,230, 23%) and Australia/New Zealand (n = 1,335, 23%) were idle. PIVC documentation in hospital records was also poor, nearly half of all PIVCs (n = 19,768, 49%) had no documented date and time of insertion. The poorest compliance was in Australia/New Zealand (n = 3,428, 59%; Appendix 4.6). We also observed that 1 in 10 PIVCs had no documentation regarding who inserted the PIVC (n = 3,905). Thirty-six percent of PIVCs (n = 14,787) had no documented assessment of the PIVC site on the day of review (Appendix 4.7), including over half of all PIVCs in Asia (n = 3,364, 52%). Overall, the median dwell at the time of assessment for PIVCs with insertion date/time documented was 1.5 d (IQR, 1.0–2.5 d).

 

DISCUSSION

This international assessment of more than 40,000 PIVCs in 51 countries provides great insight into device characteristics and variation in management practices. Predominantly, PIVCs were inserted by nurses in the general ward environment for IV medication. One in ten PIVCs had at least one symptom of phlebitis, one in ten were dysfunctional, one in five PIVC dressings were compromised, and one in six PIVCs had not been used in the preceding 24 h. Nearly half of the PIVCs audited had the insertion date and time missing.

Regional variation was found in the professions inserting PIVCs, as well as in anatomical placement. In Australia/New Zealand, the proportion of nurses inserting PIVCs was much lower than the study group average (26% vs 71%). Because these countries contributed a substantial number of hospitals to the study, this seems a representative finding and suggests a need for education targeted at nurses for PIVC insertion in this region. The veins in the forearm are recommended as optimal for PIVC insertion in adults, rather than areas of high flexion, because the forearm provides a wide surface area to secure and dress PIVCs. Forearm placement can reduce pain during catheter dwell as well as decrease the risk of accidental removal or occlusion.^3,19,27 We found only one-third of PIVCs were placed in the forearm, with most placed in the hand, antecubital veins, or wrist. This highlights an inconsistency with published recommendations and suggests that additional training and technology are required so that staff can better identify and insert PIVCs in the forearm for other than very short-term (procedural) PIVCp;s.¹⁹

Phlebitis triggering PIVC failure remains a global clinical challenge with numerous phlebitis definitions and varied assessment techniques.¹⁰ The prevalence of phlebitis has been difficult to approximate with varying estimates and definitions in the literature; however, it remains a key predictor of PIVC failure.^6,10 Identification of this complication and prompt removal of the device is critical for patient comfort and reducing CABSI risk.^5,28 The overall prevalence of phlebitis signs or symptoms (defined in this study as having one or more signs of redness, swelling, or pain surrounding the insertion site) was just over 10%, with pain and/or redness being most prevalent. These compromised PIVCs had not been removed as is recommended for such complications.^19,28 Considering that our study was a snapshot at only one time point, the per-catheter incidence of phlebitis would be even higher; interestingly, among PIVCs with a documented insertion date and time, we observed that dwell time did not influence phlebitis rates.

Another concern is that nearly 10% (n = 3,879) of PIVCs were malfunctioning (eg, leaking) but were still in place. To bring these problems into context, around 2 billion PIVCs are used annually worldwide; as a consequence, millions of patients suffer from painful or malfunctioning PIVCs staff had not responded.^1,29 The placement of large-bore catheters, and smaller-gauge ones in adults, is known to increase the incidence of malfunction that leads to failure. There are a number of sound clinical reasons for the use of large-bore (eg, resuscitation and rapid fluid replacement) or small-bore (eg, difficult venous access with small superficial veins only visible and palpable) catheters. However, it would be expected that only a small proportion of patients would require these devices, and not one in three devices as we identified. This finding suggests that some PIVCs were inappropriate in size for general IV therapy and may reflect antiquated hospital policies for some clinical cohorts.^30,31

Overall, transparent dressings were used to cover the PIVC, but a number of patients were observed to have a sterile gauze and tape dressing (n = 2,592, 6%). Although the latter is less common, both dressing approaches are recommended in clinical practice guidelines because there is a lack of high-quality evidence regarding which is superior.^21,22,32 Of concern was the use of nonsterile tape to dress the PIVC (n = 5,169, 12.7%). We found the prevalence of nonsterile tape use to be higher in lower-resourced countries in South America (n = 714, 30%), Africa (n = 543, 19%) and Europe (n = 3,056, 18%) and this was likely related to institutional cost reduction practices.

This finding illustrates an important issue regarding proper PIVC care and management practices in developing nations. It is widely known that access to safe health care in lower-resourced nations is challenging and that rates of mortality related to healthcare-associated infections are much higher. Thus, the differences we found in PIVC management practices in these countries are not surprising.^33,34 International health networks such as the Infection Control Africa Network, the International Federation of Infection Control, and the Centers for Disease Control and Prevention can have great influence on ministries of health and clinicians in these countries to develop coordinated efforts for safe and sustainable IV practices to reduce the burden of hospital-acquired infections and related morbidity and mortality.

We found that 14% of all PIVCs had no documented IV medication or IV fluid administered in the previous 24 h, strongly indicating that they were no longer needed. Australia/New Zealand, Europe, and North America were observed to have a higher prevalence of idle catheters than the remaining regions. This suggests that an opportunity exists to develop surveillance systems that better identify idle devices for prompt removal to reduce infection risk and patient discomfort. Several randomized controlled trials, a Cochrane review, and clinical practice guidelines recommend prompt removal of PIVCs when not required, if there are any complications, or if the PIVC was inserted urgently without an aseptic insertion technique.^21,28,35,36 Idle PIVCs have been implicated in adverse patient outcomes, including phlebitis and CABSI.^13,27

The substantial proportion of patients with a PIVC in this study who had no clinical indication for a PIVC, a symptomatic insertion site, malfunctioning catheter, and suboptimal dressing quality suggests the need for physicians, advanced practitioners, and nurses to adopt evidence-based PIVC insertion and maintenance bundles and supporting checklists to reduce the prevalence of PIVC complications.^19,21,38-40 Recommended strategies for inclusion in PIVC maintenance bundles are prompt removal of symptomatic and/or idle catheters, hand hygiene prior to accessing the catheter, regular assessment of the device, and replacement of suboptimal dressings.^41,42 This approach should be implemented across all clinical specialties involved in PIVC insertion and care.

Our study findings need to be considered within the context of some limitations. The cross-sectional design prevented follow-up of PIVCs until removal to collect outcomes, including subsequent PIVC complications and/or failure, following the study observation. Ideally, data collection could have included patient-level preferences for PIVC insertion, history of PIVC use and/or failure, the number of PIVC insertion attempts, and the number of PIVCs used during that hospitalization. However, a cohort study of this magnitude was not feasible, particularly because all sites contributed staff time to complete the data collection. Only half of all initially registered sites eventually participated in the study; reasons for not participating were cited as local workload constraints and/or difficulties in applying for local approvals. Although efforts to enroll hospitals worldwide were exhaustive, our sample was not randomly selected but relied on self-selection and so is not representative, particularly for countries that contributed only one hospital site. Caution is also required when comparing inter regional differences, particularly developing regions, because better-resourced/academic sites were possibly over represented in the sample. Nevertheless, PIVC variables differed significantly between participating hospitals, suggesting that the data represent a reasonable reflection of hospital variability.

 

CONCLUSIONS

On the basis of this international investigation, we report variations in the characteristics, management practices, and outcomes of PIVCs inserted in hospital patients from 51 countries. Many PIVCs were idle, symptomatic, had substandard dressings, and were inserted in suboptimal anatomical sites. Despite international best practice guidelines, a large number of patients had PIVCs that were already failing or at risk of complications, including infection. A stronger focus is needed on compliance with PIVC insertion and management guidelines; better surveillance of PIVC sites; and improved assessment, decision-making, and documentation.

Acknowledgements

We are extremely grateful to colleagues from across the globe who committed their time and effort to this study (for full details of countries and team members see Appendix 1).

The majority of hospitalized patients worldwide have at least one peripheral intravenous catheter (PIVC),¹ making PIVC insertion one of the most common clinical procedures. In the United States, physicians, advanced practitioners, and nurses insert over 300 million of these devices in hospitalized patients annually.² Despite their prevalence, PIVCs are associated with high rates of complications, including insertion difficulty, phlebitis, infiltration, occlusion, dislodgment, and catheter-associated bloodstream infection (CABSI), known to increase morbidity and mortality risk.^2-9 Up to 90% of PIVCs are prematurely removed owing to failure before planned replacement or before intravenous (IV) therapy completion.^3-6,10-12

PIVC complication and failure commonly triggers insertion of a replacement device and can entail significant costs.^2-4 One example is PIVC-related CABSI, where treatment costs have been estimated to be between US$35,000 and US$56,000 per patient.^6,13 Another important consideration is the pain and anxiety experienced by patients who need a replacement device, particularly those with difficult vascular access, who may require multiple cannulation attempts to replace a PIVC.^12,14-16 In developing nations, serious adverse events related to PIVCs are even more concerning, because hospital acquired infection rates and associated mortality are nearly 20 times greater than in developed nations.¹⁷

A number of evidence-based interventions have been suggested to reduce PIVC failure rates. In addition to optimal hand hygiene when inserting or accessing a PIVC to prevent infection,¹⁸ recommended interventions include placement of the PIVC in an area of non-flexion such as the forearm to provide stability for the device and to reduce patient discomfort, securing the PIVC to reduce movement of the catheter at the insertion site and within the blood vessel, and use of occlusive dressings that reduce the risk of external contamination of the PIVC site.^11,19,20 Best practice guidelines also recommend the prompt removal of devices that are symptomatic (when phlebitis or other complications are suspected) and when the catheter is no longer required.^21,22

Recent evidence has demonstrated that catheter size can have an impact on device survival rates. In adults, large-bore catheters of 18 gauge (G) or higher were found to have an increased rate of thrombosis, and smaller-bore catheters of 22G or lower (in adults) were found to have higher rates of dislodgment and occlusion/infiltration. The catheter size recommended for adults based on the latest evidence for most clinical applications is 20G.^3,20,23,24 In addition, the documentation of insertion, maintenance, and removal of PIVCs in the medical record is a requirement in most healthcare facilities worldwide and is recommended by best practice guidelines; however, adherence remains a challenge.^1,19

The concerning prevalence of PIVC-related complications and the lack of comparative data internationally on organizational compliance with best practice guidelines formed the rationale for this study. Our study aim was to describe the insertion characteristics, management practices, and outcomes of PIVCs internationally and to compare these variables to recommended best practice.

MATERIALS AND METHODS

Study Design and Participants

In this international cross-sectional study, we recruited hospitals through professional networks, including vascular access, infection prevention, safety and quality, nursing, and hospital associations (Appendix 2). Healthcare organizations, government health departments, and intravascular device suppliers were informed of the study and requested to further disseminate information through their networks. A study website was developed,²⁵ and social media outlets, including Twitter^®, LinkedIn^®, and Facebook^®, were used to promote the study.

 

Approval was granted by the Griffith University Human Research Ethics Committee in Australia (reference number NRS/34/13/HREC). In addition, evidence of study site and local institutional review board/ethics committee approval was required prior to study commencement. Each participating site agreed to follow the study protocol and signed an authorship agreement form. No financial support was provided to any site.

Hospitalized adult and pediatric patients with a PIVC in situ on the day of the study were eligible for inclusion. Sample size was determined by local capacity. Hospitals were encouraged to audit their entire institution if possible; however, data were accepted from as little as one ward. Data collectors comprised nurses and doctors with experience in PIVC assessment. They were briefed on the study protocol and data collection forms by the local site coordinator, and they were supported by an overall global coordinator. Clinicians assessed the PIVC insertion site and accessed hospital records to collect data related to PIVC insertion, concurrent medications, and IV fluid orders. Further clarification of data was obtained if necessary by the clinicians from the patients and treating staff. No identifiable patient information was collected.

Data Collection

To assess whether clinical facilities were following best practice recommendations, the study team developed three data collection forms to collect information regarding site characteristics (site questionnaire), track participant recruitment (screening log), and collect data regarding PIVC characteristics and management practices (case report form [CRF]). All forms were internally and externally validated following a pilot study involving 14 sites in 13 countries.¹

The CRF included variables used to assess best practice interventions, such as catheter insertion characteristics (date and time, reason, location, profession of inserter, anatomical site of placement), catheter type (gauge, brand, and product), insertion site assessment (adverse symptoms, dressing type and integrity), and information related to the IV therapy (types of IV fluids and medications, flushing solutions). Idle PIVCs were defined as not being used for blood sampling or IV therapy in the preceding 24 h.

Data collection forms were translated into 15 languages by professional translators and back-translated for validity. Translation of some languages included additional rigor. For example, Spanish-speaking members from the Spanish mainland as well as from South America were employed so that appropriate synonyms were used to capture local terms and practice. Three options were provided for data entry: directly into a purpose-developed electronic database (Lime Survey^® Project, Hamburg, Germany); on paper, then transcribed into the survey database at a later time by the hospital site; or paper entry then sent (via email or post) to the coordinating center for data entry. Once cleaned and collated, all data were provided to each participating hospital to confirm accuracy and for site use in local quality improvement processes. Data were collected between June 1, 2014 and July 31, 2015.

Statistical Analysis

All data management was undertaken using SAS statistical software (SAS Institute Inc., Cary NC, USA). Results are presented for eight geographical regions using descriptive statistics (frequencies, percentages, and 95% CIs) for the variables of interest. To assess trends in catheter dwell time and rates of phlebitis, Poisson regression was used. All analyses were undertaken using the R language for statistical analysis (R Core Team, Vienna, Austria). The (STROBE (Strengthening the Reporting of Observational Studies in Epidemiology statement) guidelines for cross-sectional studies were followed, and results are presented according to these recommendations.²⁶

RESULTS

Of the 415 hospitals that participated in this study, 406 had patients with PIVCs on the day of the study (the others being small rural centers). Thus, a total of 40,620 PIVCs in 38,161 patients from 406 hospitals in 51 countries were assessed, with no more than 5% missing data for any CRF question. There were 2459 patients (6.1%) with two or more PIVCs concurrently in situ. The median patient age was 59 y (interquartile range [IQR], 37–74 y), and just over half were male (n = 20,550, 51%). Hospital size ranged from fewer than 10 beds to over 1,000 beds, and hospitals were located in rural, regional, and metropolitan districts. The majority of countries (n = 31, 61%) contributed multiple sites, the highest being Australia with 79 hospitals. Countries with the most PIVCs studied were Spain (n = 5,553, 14%) and the United States (n = 5,048, 12%).

General surgical (n = 15,616, 39%) and medical (n = 15,448, 38%) patients represented most of the population observed. PIVCs were inserted primarily in general wards or clinics (n = 22,167, 55%) or in emergency departments (n = 7,388, 18%; Table) and for the administration of IV medication (n = 28,571, 70%) and IV fluids (n = 7,093, 18%; Table).

Globally, nurses were the primary PIVC inserters (n = 28,575, 71%); however, Australia/New Zealand had only 26% (n = 1,518) of PIVCs inserted by this group (Table). Only about one-third of PIVCs were placed in an area of non-flexion (forearm, n = 12,675, 31%, Table) the majority (n = 27,856, 69%) were placed in non-recommended anatomical sites (Figure 1). Most PIVCs were placed in the hand (n = 13,265, 32.7%) followed by the antecubital veins (n = 6176, 15.2%) and the wrist (n = 5,465, 13.5%). Site selection varied widely across the regions; 29% (n = 1686) of PIVCs in Australia/New Zealand were inserted into the antecubital veins, twice the study group average. Over half of the PIVCs inserted in the Middle East were placed in the hand (n = 295, 56%). This region also had the highest prevalence of devices placed in nonrecommended sites (n = 416, 79%; Figure 1).

The majority of PIVCs (n = 27,192, 67%; Table) were of recommended size (20–22G); however, some devices were observed to be large (14–18G; n = 6,802, 17%) or small (24-26g; n = 4,869, 12%) in adults. In Asia, 41% (n = 2,617) of devices inserted were 24-26G, more than three times the global rate. Half of all devices in Asia (n = 3,077, 48%) and the South Pacific (n = 67, 52%) were of a size not recommended for routine IV therapy (Figure 2).

The primary dressing material used was a transparent dressing (n = 31,596, 77.8%; Table); however, nearly 1 in 5 dressings used had either nonsterile tape alone (n = 5,169, 13%; Appendix 4), or a sterile gauze and tape (n = 2,592, 6%; Appendix 4.1). We found a wide variation in the use of nonsterile tape, including 1 in every 3 devices in South America dressed with nonsterile tape (n = 714, 30%) and a larger proportion in Africa (n = 543, 19%) and Europe (n = 3,056, 18%). Nonsterile tape was rarely used in North America and Australia/New Zealand. Although most PIVC dressings were clean, dry, and intact (n = 31,786, 79%; Table), one-fifth overall were compromised (moist, soiled, and/or lifting off the skin). Compromised dressings (Appendix 4.2) were more prevalent in Australia/New Zealand (n = 1,448; 25%) and in Africa (n = 707, 25%) than elsewhere.

Ten percent of PIVCs (n = 4,204) had signs and/or symptoms suggestive of phlebitis (characterized by pain, redness and/or swelling at the insertion site; Appendix 4.3). The highest prevalence of phlebitis occurred in Asia (n = 1,021, 16%), Africa (n = 360, 13%), and South America (n = 284, 12%). Pain and/or redness were the most common phlebitis symptoms. We found no association between dwell time of PIVCs and phlebitis rates (P = .085). Phlebitis rates were 12% (Days 1-3; n = 15,625), 16% (Days 4-7; n = 3,348), 10% (Days 8-21; n = 457), and 13% (Day21+; n = 174). Nearly 10% (n = 3,879) of catheters were observed to have signs of malfunction such as blood in the infusion tubing, leaking at the insertion site, or dislodgment (Appendix 4.4).

We observed 14% (n = 5,796) of PIVCs to be idle (Appendix 4.5), defined as not used in the preceding 24 h. Nearly one-fourth of all devices in North America (n = 1,230, 23%) and Australia/New Zealand (n = 1,335, 23%) were idle. PIVC documentation in hospital records was also poor, nearly half of all PIVCs (n = 19,768, 49%) had no documented date and time of insertion. The poorest compliance was in Australia/New Zealand (n = 3,428, 59%; Appendix 4.6). We also observed that 1 in 10 PIVCs had no documentation regarding who inserted the PIVC (n = 3,905). Thirty-six percent of PIVCs (n = 14,787) had no documented assessment of the PIVC site on the day of review (Appendix 4.7), including over half of all PIVCs in Asia (n = 3,364, 52%). Overall, the median dwell at the time of assessment for PIVCs with insertion date/time documented was 1.5 d (IQR, 1.0–2.5 d).

 

DISCUSSION

This international assessment of more than 40,000 PIVCs in 51 countries provides great insight into device characteristics and variation in management practices. Predominantly, PIVCs were inserted by nurses in the general ward environment for IV medication. One in ten PIVCs had at least one symptom of phlebitis, one in ten were dysfunctional, one in five PIVC dressings were compromised, and one in six PIVCs had not been used in the preceding 24 h. Nearly half of the PIVCs audited had the insertion date and time missing.

Regional variation was found in the professions inserting PIVCs, as well as in anatomical placement. In Australia/New Zealand, the proportion of nurses inserting PIVCs was much lower than the study group average (26% vs 71%). Because these countries contributed a substantial number of hospitals to the study, this seems a representative finding and suggests a need for education targeted at nurses for PIVC insertion in this region. The veins in the forearm are recommended as optimal for PIVC insertion in adults, rather than areas of high flexion, because the forearm provides a wide surface area to secure and dress PIVCs. Forearm placement can reduce pain during catheter dwell as well as decrease the risk of accidental removal or occlusion.^3,19,27 We found only one-third of PIVCs were placed in the forearm, with most placed in the hand, antecubital veins, or wrist. This highlights an inconsistency with published recommendations and suggests that additional training and technology are required so that staff can better identify and insert PIVCs in the forearm for other than very short-term (procedural) PIVCp;s.¹⁹

Phlebitis triggering PIVC failure remains a global clinical challenge with numerous phlebitis definitions and varied assessment techniques.¹⁰ The prevalence of phlebitis has been difficult to approximate with varying estimates and definitions in the literature; however, it remains a key predictor of PIVC failure.^6,10 Identification of this complication and prompt removal of the device is critical for patient comfort and reducing CABSI risk.^5,28 The overall prevalence of phlebitis signs or symptoms (defined in this study as having one or more signs of redness, swelling, or pain surrounding the insertion site) was just over 10%, with pain and/or redness being most prevalent. These compromised PIVCs had not been removed as is recommended for such complications.^19,28 Considering that our study was a snapshot at only one time point, the per-catheter incidence of phlebitis would be even higher; interestingly, among PIVCs with a documented insertion date and time, we observed that dwell time did not influence phlebitis rates.

Another concern is that nearly 10% (n = 3,879) of PIVCs were malfunctioning (eg, leaking) but were still in place. To bring these problems into context, around 2 billion PIVCs are used annually worldwide; as a consequence, millions of patients suffer from painful or malfunctioning PIVCs staff had not responded.^1,29 The placement of large-bore catheters, and smaller-gauge ones in adults, is known to increase the incidence of malfunction that leads to failure. There are a number of sound clinical reasons for the use of large-bore (eg, resuscitation and rapid fluid replacement) or small-bore (eg, difficult venous access with small superficial veins only visible and palpable) catheters. However, it would be expected that only a small proportion of patients would require these devices, and not one in three devices as we identified. This finding suggests that some PIVCs were inappropriate in size for general IV therapy and may reflect antiquated hospital policies for some clinical cohorts.^30,31

Overall, transparent dressings were used to cover the PIVC, but a number of patients were observed to have a sterile gauze and tape dressing (n = 2,592, 6%). Although the latter is less common, both dressing approaches are recommended in clinical practice guidelines because there is a lack of high-quality evidence regarding which is superior.^21,22,32 Of concern was the use of nonsterile tape to dress the PIVC (n = 5,169, 12.7%). We found the prevalence of nonsterile tape use to be higher in lower-resourced countries in South America (n = 714, 30%), Africa (n = 543, 19%) and Europe (n = 3,056, 18%) and this was likely related to institutional cost reduction practices.

This finding illustrates an important issue regarding proper PIVC care and management practices in developing nations. It is widely known that access to safe health care in lower-resourced nations is challenging and that rates of mortality related to healthcare-associated infections are much higher. Thus, the differences we found in PIVC management practices in these countries are not surprising.^33,34 International health networks such as the Infection Control Africa Network, the International Federation of Infection Control, and the Centers for Disease Control and Prevention can have great influence on ministries of health and clinicians in these countries to develop coordinated efforts for safe and sustainable IV practices to reduce the burden of hospital-acquired infections and related morbidity and mortality.

We found that 14% of all PIVCs had no documented IV medication or IV fluid administered in the previous 24 h, strongly indicating that they were no longer needed. Australia/New Zealand, Europe, and North America were observed to have a higher prevalence of idle catheters than the remaining regions. This suggests that an opportunity exists to develop surveillance systems that better identify idle devices for prompt removal to reduce infection risk and patient discomfort. Several randomized controlled trials, a Cochrane review, and clinical practice guidelines recommend prompt removal of PIVCs when not required, if there are any complications, or if the PIVC was inserted urgently without an aseptic insertion technique.^21,28,35,36 Idle PIVCs have been implicated in adverse patient outcomes, including phlebitis and CABSI.^13,27

The substantial proportion of patients with a PIVC in this study who had no clinical indication for a PIVC, a symptomatic insertion site, malfunctioning catheter, and suboptimal dressing quality suggests the need for physicians, advanced practitioners, and nurses to adopt evidence-based PIVC insertion and maintenance bundles and supporting checklists to reduce the prevalence of PIVC complications.^19,21,38-40 Recommended strategies for inclusion in PIVC maintenance bundles are prompt removal of symptomatic and/or idle catheters, hand hygiene prior to accessing the catheter, regular assessment of the device, and replacement of suboptimal dressings.^41,42 This approach should be implemented across all clinical specialties involved in PIVC insertion and care.

Our study findings need to be considered within the context of some limitations. The cross-sectional design prevented follow-up of PIVCs until removal to collect outcomes, including subsequent PIVC complications and/or failure, following the study observation. Ideally, data collection could have included patient-level preferences for PIVC insertion, history of PIVC use and/or failure, the number of PIVC insertion attempts, and the number of PIVCs used during that hospitalization. However, a cohort study of this magnitude was not feasible, particularly because all sites contributed staff time to complete the data collection. Only half of all initially registered sites eventually participated in the study; reasons for not participating were cited as local workload constraints and/or difficulties in applying for local approvals. Although efforts to enroll hospitals worldwide were exhaustive, our sample was not randomly selected but relied on self-selection and so is not representative, particularly for countries that contributed only one hospital site. Caution is also required when comparing inter regional differences, particularly developing regions, because better-resourced/academic sites were possibly over represented in the sample. Nevertheless, PIVC variables differed significantly between participating hospitals, suggesting that the data represent a reasonable reflection of hospital variability.

 

CONCLUSIONS

On the basis of this international investigation, we report variations in the characteristics, management practices, and outcomes of PIVCs inserted in hospital patients from 51 countries. Many PIVCs were idle, symptomatic, had substandard dressings, and were inserted in suboptimal anatomical sites. Despite international best practice guidelines, a large number of patients had PIVCs that were already failing or at risk of complications, including infection. A stronger focus is needed on compliance with PIVC insertion and management guidelines; better surveillance of PIVC sites; and improved assessment, decision-making, and documentation.

Acknowledgements

We are extremely grateful to colleagues from across the globe who committed their time and effort to this study (for full details of countries and team members see Appendix 1).

Head and neck cancer (HNC) continues to be a major health issue with an estimated 51,540 cases in the US in 2018, making it the eighth most common cancer among men with an estimated 4% of all new cancer diagnoses.¹ Over the past decade, human papillomavirus (HPV) has emerged as a major prognostic factor for survival in squamous cell carcinomas of the oropharynx. Patients who are HPV-positive (HPV+) have a much higher survival rate than patients who have HPV-negative (HPV-) cancers of the oropharynx. The 8th edition of the American Joint Committee on Cancer (AJCC) staging manual has 2 distinct stagings for HPV+ and HPV- oropharyngeal tumors using p16-positivity (p16+) as a surrogate marker.²

Squamous cell carcinomas of the oropharynx that are HPV+ have about half the risk of death of HPV- tumors, are highly responsive to treatment, and are more often seen in younger and healthier patients with little to no tobacco use.^2,3 As such, there also is a movement to de-escalate HPV+ oropharyngeal cancers with multiple trials by either replacing cytotoxic chemotherapy with a targeted agent (cisplatin vs cetuximab in RTOG 1016) or reducing the radiation dose (ECOG 1308, NRG HN002, Quarterback, and OPTIMA trials).³

The focus of many epidemiologic studies has been in the HNC general population. A recent epidemiologic analysis of the HNC general population found a p16 positivity rate of 60% in oropharyngeal squamous cell carcinomas (OPSCC) and 10% in nonoropharyngeal squamous cell carcinomas (NOPSCC).⁴ There has been a lack of studies focusing on the US Department of Veterans Administration (VA) population. The VA HNC population consists mostly of older white male smokers; whereas the rise of OPSCC in the general population consists primarily of males aged < 60 years often with little or no tobacco use.⁵ Furthermore, the importance of p16 positivity in NOPSCC also may be prognostic.⁶ Population data on this subset in the VA are lacking as well.This study’s purpose is to analyze the p16 positivity rate in both the OPSCC and NOPSCC in the VA population. Elucidation of epidemiologic factors that are associated with these groups may bring to light important differences between the VA and general HNC populations.

Methods

A review of the Kansas City VA Medical Center database for patients with HNC was performed from 2011 to 2017. The review consisted of 183 patient records (second primaries were scored separately), and 123 were deemed eligible for the study. Epidemiologic data were collected, including site, OPSCC vs NOPSCC, age, race, education level, tobacco use, alcohol use, TNM stage, and marital status (Table). 

Gender was not included because there was only 1 female patient in the cohort. Four subgroups based on site and p16 status (OPSCC p16+, OPSCC p16-, NOPSCC p16+, and NOPSCC p16-) were further analyzed. Appropriate statistical analysis (chi-square test, analysis of variance, and Kruskal-Wallis test) with IBM SPSS 24.0 (Armonk, NY) was used to find differences (P < .05) among the means of the 4 subgroups.

Results

There were 55 (44%) patients with OPSCC and 68 patients with NOPSCC (56%). Of the 68 patients with NOPSCC, 48 (70%) were primary tumors from the larynx, 12 (18%) from the oral cavity, 4 (6%) from the hypopharynx, 2 from the nasopharynx (3%), and 2 (3%) were unknown primaries. In the OPSCC group, 41 patients were p16+ (75%) and 14 p16- (25%). In the NOPSCC group, 20 patients were p16+ (29%) and 48 were p16- (71%). There was a statistically significant difference seen in tobacco use, TNM stage, and marital status. Alcohol use trended toward significance.

The NOPSCC p16+ group had the greatest mean pack-year use (57). The lowest was in the OPSCC p16+ group (29). The OPSCC p16+ group had 37% never smokers compared with ≤ 10% for the other groups. Both the OPSCC and NOPSCC p16- groups had much more alcohol use per week than that of the p16+ groups. The differences in marital status included a lower rate of never married individuals in the p16+ group and a higher rate of marriage in the NOPSCC p16- group. The T stage distribution within the OPSCC groups was similar, but NOPSCC groups saw more T1 lesions in the NOPSCC p16- group (42% p16- vs 18% p16+). Conversely, more T4 lesions were found in the NOPSCC p16+ patients (7% p16- vs 29% p16+). More advanced nodal staging was seen in both OPSCC groups with 78% N2 or N3 in the p16+ group and 82% in the p16- group. The NOPSCC p16+ group had 55% N0 or N1 patients, and the p16- group had 60%. In terms of overall stage, the OPSCC groups had a similar distribution with predominantly stage IVA/B presentation (74% p16+ and 82% p16-), whereas the NOPSCC groups had only 58% (p16+) and 35% (p16-) at presentation.

Discussion

The overall HPV positivity rate in the general population of patients with HNC has been reported as between 57% and 72% for OPSCC and between 1.3% and 7% for NOPSCC.⁶ One study, however, examined the p16 positivity rate in NOPSCC patients enrolled in major trials (RTOG 0129, 0234, and 0522 studies) and found that up to 19.3% of NOPSCC patients had p16 positivity.⁶ Even with the near 20% rate in those aforementioned trials that are above the reported norm, the current study found that nearly 30% of its VA population had p16+ NOPSCC. It has been shown that regardless of site, HPV-driven head and neck tumors share a similar gene expression and DNA methylation profiles (nonkeratinizing, basaloid histopathologic features, and lack of TP53 or CDKN2A alterations).⁵ p16+ NOPSCC has a different immune microenvironment with less lymphocyte infiltration, and there is some debate in the literature about the effects on tumor outcomes for NOPSCC cancer.⁵

In the aforementioned RTOG trials, p16- NOPSCC had worse outcomes compared with those of p16+ NOPSCC.⁶ This result is in contrast to the Danish Head and Neck Cancer Group (DAHANCA) and the combined Johns Hopkins University (JHU) and University of California, San Francisco (UCSF) data that found no difference between p16+ NOPSCC or p16- NOPSCC.^7,8 In regards to race, this study did not find any differences. Another UCSF and JHU study showed lower p16+ rates in African American patients with OPSCC, but no distinction between race in the NOPSCC group. This result is consistent with the data in the current study as the distribution of race was no different among the 4 groups; however, this study's cohort was 90% white, 10% African American, and only < 1% Native American.⁴ This study's cohort population also was consistent with HPV-positive tumors presenting with earlier T, but higher N staging.⁹

Smoking is known to decrease survival in HPV-positive HNC, with the RTOG 0129 study separating head and neck tumors into low, medium, and high risk, based on HPV status, smoking, and stage.¹⁰ Although the average smoking pack-years in the current study’s OPC p16+ group was high at 29 pack-years, there was still a significant number of nonsmokers in that same group (37%). The University of Michigan conducted a study that had a similar profile of patients with an average age of 56.5 and 32.4% never smokers in their p16+ OPSCC cohort; thus, the VA p16+ OPSCC group in this study may be similar to the general population's p16+ OPSCC group.¹¹ Nonmonogamous relationships also have been shown to be a risk factor for HPV positivity, and there was a difference in marital status (assuming it was a surrogate for monogamy) between the 4 groups; however, in contrast, the p16+ group in the current study had a high number of married patients, 45% in OPC p16+ group, and may not have been a good surrogate for monogamy in this VA population.

Limitations

Limitations of this study include all the caveats that come with a retrospective study, such as confounding variables, unbalanced groups, and selection bias. A detailed sexual history was not included, although it is well known that sexual activity is linked with oral HPV positivity.¹² Human papillomavirus positivity based on p16 immunohistochemical analysis also was used as a surrogate marker for HPV instead of DNA in situ hybridization. The data also may be skewed due to the study patient’s being predominantly white and male: Both groups have a higher predilection for HPV-driven HNCs.¹³

Conclusion

The proportion of p16+ VA OPSCC cases was similar to that of the general population at 75% with 37% never smokers, but the percentage in NOPSCC was higher at 29% with only 10% never smokers. The p16+ NOPSCC also presented with more T4 lesions and a higher overall stage compared with p16- NOPSCC. Further studies are needed to compare these subgroups in the VA and in the general HNC populations.

Head and neck cancer (HNC) continues to be a major health issue with an estimated 51,540 cases in the US in 2018, making it the eighth most common cancer among men with an estimated 4% of all new cancer diagnoses.¹ Over the past decade, human papillomavirus (HPV) has emerged as a major prognostic factor for survival in squamous cell carcinomas of the oropharynx. Patients who are HPV-positive (HPV+) have a much higher survival rate than patients who have HPV-negative (HPV-) cancers of the oropharynx. The 8th edition of the American Joint Committee on Cancer (AJCC) staging manual has 2 distinct stagings for HPV+ and HPV- oropharyngeal tumors using p16-positivity (p16+) as a surrogate marker.²

Squamous cell carcinomas of the oropharynx that are HPV+ have about half the risk of death of HPV- tumors, are highly responsive to treatment, and are more often seen in younger and healthier patients with little to no tobacco use.^2,3 As such, there also is a movement to de-escalate HPV+ oropharyngeal cancers with multiple trials by either replacing cytotoxic chemotherapy with a targeted agent (cisplatin vs cetuximab in RTOG 1016) or reducing the radiation dose (ECOG 1308, NRG HN002, Quarterback, and OPTIMA trials).³

The focus of many epidemiologic studies has been in the HNC general population. A recent epidemiologic analysis of the HNC general population found a p16 positivity rate of 60% in oropharyngeal squamous cell carcinomas (OPSCC) and 10% in nonoropharyngeal squamous cell carcinomas (NOPSCC).⁴ There has been a lack of studies focusing on the US Department of Veterans Administration (VA) population. The VA HNC population consists mostly of older white male smokers; whereas the rise of OPSCC in the general population consists primarily of males aged < 60 years often with little or no tobacco use.⁵ Furthermore, the importance of p16 positivity in NOPSCC also may be prognostic.⁶ Population data on this subset in the VA are lacking as well.This study’s purpose is to analyze the p16 positivity rate in both the OPSCC and NOPSCC in the VA population. Elucidation of epidemiologic factors that are associated with these groups may bring to light important differences between the VA and general HNC populations.

Methods

A review of the Kansas City VA Medical Center database for patients with HNC was performed from 2011 to 2017. The review consisted of 183 patient records (second primaries were scored separately), and 123 were deemed eligible for the study. Epidemiologic data were collected, including site, OPSCC vs NOPSCC, age, race, education level, tobacco use, alcohol use, TNM stage, and marital status (Table). 

Gender was not included because there was only 1 female patient in the cohort. Four subgroups based on site and p16 status (OPSCC p16+, OPSCC p16-, NOPSCC p16+, and NOPSCC p16-) were further analyzed. Appropriate statistical analysis (chi-square test, analysis of variance, and Kruskal-Wallis test) with IBM SPSS 24.0 (Armonk, NY) was used to find differences (P < .05) among the means of the 4 subgroups.

Results

There were 55 (44%) patients with OPSCC and 68 patients with NOPSCC (56%). Of the 68 patients with NOPSCC, 48 (70%) were primary tumors from the larynx, 12 (18%) from the oral cavity, 4 (6%) from the hypopharynx, 2 from the nasopharynx (3%), and 2 (3%) were unknown primaries. In the OPSCC group, 41 patients were p16+ (75%) and 14 p16- (25%). In the NOPSCC group, 20 patients were p16+ (29%) and 48 were p16- (71%). There was a statistically significant difference seen in tobacco use, TNM stage, and marital status. Alcohol use trended toward significance.

The NOPSCC p16+ group had the greatest mean pack-year use (57). The lowest was in the OPSCC p16+ group (29). The OPSCC p16+ group had 37% never smokers compared with ≤ 10% for the other groups. Both the OPSCC and NOPSCC p16- groups had much more alcohol use per week than that of the p16+ groups. The differences in marital status included a lower rate of never married individuals in the p16+ group and a higher rate of marriage in the NOPSCC p16- group. The T stage distribution within the OPSCC groups was similar, but NOPSCC groups saw more T1 lesions in the NOPSCC p16- group (42% p16- vs 18% p16+). Conversely, more T4 lesions were found in the NOPSCC p16+ patients (7% p16- vs 29% p16+). More advanced nodal staging was seen in both OPSCC groups with 78% N2 or N3 in the p16+ group and 82% in the p16- group. The NOPSCC p16+ group had 55% N0 or N1 patients, and the p16- group had 60%. In terms of overall stage, the OPSCC groups had a similar distribution with predominantly stage IVA/B presentation (74% p16+ and 82% p16-), whereas the NOPSCC groups had only 58% (p16+) and 35% (p16-) at presentation.

Discussion

The overall HPV positivity rate in the general population of patients with HNC has been reported as between 57% and 72% for OPSCC and between 1.3% and 7% for NOPSCC.⁶ One study, however, examined the p16 positivity rate in NOPSCC patients enrolled in major trials (RTOG 0129, 0234, and 0522 studies) and found that up to 19.3% of NOPSCC patients had p16 positivity.⁶ Even with the near 20% rate in those aforementioned trials that are above the reported norm, the current study found that nearly 30% of its VA population had p16+ NOPSCC. It has been shown that regardless of site, HPV-driven head and neck tumors share a similar gene expression and DNA methylation profiles (nonkeratinizing, basaloid histopathologic features, and lack of TP53 or CDKN2A alterations).⁵ p16+ NOPSCC has a different immune microenvironment with less lymphocyte infiltration, and there is some debate in the literature about the effects on tumor outcomes for NOPSCC cancer.⁵

In the aforementioned RTOG trials, p16- NOPSCC had worse outcomes compared with those of p16+ NOPSCC.⁶ This result is in contrast to the Danish Head and Neck Cancer Group (DAHANCA) and the combined Johns Hopkins University (JHU) and University of California, San Francisco (UCSF) data that found no difference between p16+ NOPSCC or p16- NOPSCC.^7,8 In regards to race, this study did not find any differences. Another UCSF and JHU study showed lower p16+ rates in African American patients with OPSCC, but no distinction between race in the NOPSCC group. This result is consistent with the data in the current study as the distribution of race was no different among the 4 groups; however, this study's cohort was 90% white, 10% African American, and only < 1% Native American.⁴ This study's cohort population also was consistent with HPV-positive tumors presenting with earlier T, but higher N staging.⁹

Smoking is known to decrease survival in HPV-positive HNC, with the RTOG 0129 study separating head and neck tumors into low, medium, and high risk, based on HPV status, smoking, and stage.¹⁰ Although the average smoking pack-years in the current study’s OPC p16+ group was high at 29 pack-years, there was still a significant number of nonsmokers in that same group (37%). The University of Michigan conducted a study that had a similar profile of patients with an average age of 56.5 and 32.4% never smokers in their p16+ OPSCC cohort; thus, the VA p16+ OPSCC group in this study may be similar to the general population's p16+ OPSCC group.¹¹ Nonmonogamous relationships also have been shown to be a risk factor for HPV positivity, and there was a difference in marital status (assuming it was a surrogate for monogamy) between the 4 groups; however, in contrast, the p16+ group in the current study had a high number of married patients, 45% in OPC p16+ group, and may not have been a good surrogate for monogamy in this VA population.

Limitations

Limitations of this study include all the caveats that come with a retrospective study, such as confounding variables, unbalanced groups, and selection bias. A detailed sexual history was not included, although it is well known that sexual activity is linked with oral HPV positivity.¹² Human papillomavirus positivity based on p16 immunohistochemical analysis also was used as a surrogate marker for HPV instead of DNA in situ hybridization. The data also may be skewed due to the study patient’s being predominantly white and male: Both groups have a higher predilection for HPV-driven HNCs.¹³

Conclusion

The proportion of p16+ VA OPSCC cases was similar to that of the general population at 75% with 37% never smokers, but the percentage in NOPSCC was higher at 29% with only 10% never smokers. The p16+ NOPSCC also presented with more T4 lesions and a higher overall stage compared with p16- NOPSCC. Further studies are needed to compare these subgroups in the VA and in the general HNC populations.

Head and neck cancer (HNC) continues to be a major health issue with an estimated 51,540 cases in the US in 2018, making it the eighth most common cancer among men with an estimated 4% of all new cancer diagnoses.¹ Over the past decade, human papillomavirus (HPV) has emerged as a major prognostic factor for survival in squamous cell carcinomas of the oropharynx. Patients who are HPV-positive (HPV+) have a much higher survival rate than patients who have HPV-negative (HPV-) cancers of the oropharynx. The 8th edition of the American Joint Committee on Cancer (AJCC) staging manual has 2 distinct stagings for HPV+ and HPV- oropharyngeal tumors using p16-positivity (p16+) as a surrogate marker.²

Squamous cell carcinomas of the oropharynx that are HPV+ have about half the risk of death of HPV- tumors, are highly responsive to treatment, and are more often seen in younger and healthier patients with little to no tobacco use.^2,3 As such, there also is a movement to de-escalate HPV+ oropharyngeal cancers with multiple trials by either replacing cytotoxic chemotherapy with a targeted agent (cisplatin vs cetuximab in RTOG 1016) or reducing the radiation dose (ECOG 1308, NRG HN002, Quarterback, and OPTIMA trials).³

The focus of many epidemiologic studies has been in the HNC general population. A recent epidemiologic analysis of the HNC general population found a p16 positivity rate of 60% in oropharyngeal squamous cell carcinomas (OPSCC) and 10% in nonoropharyngeal squamous cell carcinomas (NOPSCC).⁴ There has been a lack of studies focusing on the US Department of Veterans Administration (VA) population. The VA HNC population consists mostly of older white male smokers; whereas the rise of OPSCC in the general population consists primarily of males aged < 60 years often with little or no tobacco use.⁵ Furthermore, the importance of p16 positivity in NOPSCC also may be prognostic.⁶ Population data on this subset in the VA are lacking as well.This study’s purpose is to analyze the p16 positivity rate in both the OPSCC and NOPSCC in the VA population. Elucidation of epidemiologic factors that are associated with these groups may bring to light important differences between the VA and general HNC populations.

Methods

A review of the Kansas City VA Medical Center database for patients with HNC was performed from 2011 to 2017. The review consisted of 183 patient records (second primaries were scored separately), and 123 were deemed eligible for the study. Epidemiologic data were collected, including site, OPSCC vs NOPSCC, age, race, education level, tobacco use, alcohol use, TNM stage, and marital status (Table). 

Gender was not included because there was only 1 female patient in the cohort. Four subgroups based on site and p16 status (OPSCC p16+, OPSCC p16-, NOPSCC p16+, and NOPSCC p16-) were further analyzed. Appropriate statistical analysis (chi-square test, analysis of variance, and Kruskal-Wallis test) with IBM SPSS 24.0 (Armonk, NY) was used to find differences (P < .05) among the means of the 4 subgroups.

Results

There were 55 (44%) patients with OPSCC and 68 patients with NOPSCC (56%). Of the 68 patients with NOPSCC, 48 (70%) were primary tumors from the larynx, 12 (18%) from the oral cavity, 4 (6%) from the hypopharynx, 2 from the nasopharynx (3%), and 2 (3%) were unknown primaries. In the OPSCC group, 41 patients were p16+ (75%) and 14 p16- (25%). In the NOPSCC group, 20 patients were p16+ (29%) and 48 were p16- (71%). There was a statistically significant difference seen in tobacco use, TNM stage, and marital status. Alcohol use trended toward significance.

The NOPSCC p16+ group had the greatest mean pack-year use (57). The lowest was in the OPSCC p16+ group (29). The OPSCC p16+ group had 37% never smokers compared with ≤ 10% for the other groups. Both the OPSCC and NOPSCC p16- groups had much more alcohol use per week than that of the p16+ groups. The differences in marital status included a lower rate of never married individuals in the p16+ group and a higher rate of marriage in the NOPSCC p16- group. The T stage distribution within the OPSCC groups was similar, but NOPSCC groups saw more T1 lesions in the NOPSCC p16- group (42% p16- vs 18% p16+). Conversely, more T4 lesions were found in the NOPSCC p16+ patients (7% p16- vs 29% p16+). More advanced nodal staging was seen in both OPSCC groups with 78% N2 or N3 in the p16+ group and 82% in the p16- group. The NOPSCC p16+ group had 55% N0 or N1 patients, and the p16- group had 60%. In terms of overall stage, the OPSCC groups had a similar distribution with predominantly stage IVA/B presentation (74% p16+ and 82% p16-), whereas the NOPSCC groups had only 58% (p16+) and 35% (p16-) at presentation.

Discussion

The overall HPV positivity rate in the general population of patients with HNC has been reported as between 57% and 72% for OPSCC and between 1.3% and 7% for NOPSCC.⁶ One study, however, examined the p16 positivity rate in NOPSCC patients enrolled in major trials (RTOG 0129, 0234, and 0522 studies) and found that up to 19.3% of NOPSCC patients had p16 positivity.⁶ Even with the near 20% rate in those aforementioned trials that are above the reported norm, the current study found that nearly 30% of its VA population had p16+ NOPSCC. It has been shown that regardless of site, HPV-driven head and neck tumors share a similar gene expression and DNA methylation profiles (nonkeratinizing, basaloid histopathologic features, and lack of TP53 or CDKN2A alterations).⁵ p16+ NOPSCC has a different immune microenvironment with less lymphocyte infiltration, and there is some debate in the literature about the effects on tumor outcomes for NOPSCC cancer.⁵

In the aforementioned RTOG trials, p16- NOPSCC had worse outcomes compared with those of p16+ NOPSCC.⁶ This result is in contrast to the Danish Head and Neck Cancer Group (DAHANCA) and the combined Johns Hopkins University (JHU) and University of California, San Francisco (UCSF) data that found no difference between p16+ NOPSCC or p16- NOPSCC.^7,8 In regards to race, this study did not find any differences. Another UCSF and JHU study showed lower p16+ rates in African American patients with OPSCC, but no distinction between race in the NOPSCC group. This result is consistent with the data in the current study as the distribution of race was no different among the 4 groups; however, this study's cohort was 90% white, 10% African American, and only < 1% Native American.⁴ This study's cohort population also was consistent with HPV-positive tumors presenting with earlier T, but higher N staging.⁹

Smoking is known to decrease survival in HPV-positive HNC, with the RTOG 0129 study separating head and neck tumors into low, medium, and high risk, based on HPV status, smoking, and stage.¹⁰ Although the average smoking pack-years in the current study’s OPC p16+ group was high at 29 pack-years, there was still a significant number of nonsmokers in that same group (37%). The University of Michigan conducted a study that had a similar profile of patients with an average age of 56.5 and 32.4% never smokers in their p16+ OPSCC cohort; thus, the VA p16+ OPSCC group in this study may be similar to the general population's p16+ OPSCC group.¹¹ Nonmonogamous relationships also have been shown to be a risk factor for HPV positivity, and there was a difference in marital status (assuming it was a surrogate for monogamy) between the 4 groups; however, in contrast, the p16+ group in the current study had a high number of married patients, 45% in OPC p16+ group, and may not have been a good surrogate for monogamy in this VA population.

Limitations

Limitations of this study include all the caveats that come with a retrospective study, such as confounding variables, unbalanced groups, and selection bias. A detailed sexual history was not included, although it is well known that sexual activity is linked with oral HPV positivity.¹² Human papillomavirus positivity based on p16 immunohistochemical analysis also was used as a surrogate marker for HPV instead of DNA in situ hybridization. The data also may be skewed due to the study patient’s being predominantly white and male: Both groups have a higher predilection for HPV-driven HNCs.¹³

Conclusion

The proportion of p16+ VA OPSCC cases was similar to that of the general population at 75% with 37% never smokers, but the percentage in NOPSCC was higher at 29% with only 10% never smokers. The p16+ NOPSCC also presented with more T4 lesions and a higher overall stage compared with p16- NOPSCC. Further studies are needed to compare these subgroups in the VA and in the general HNC populations.

ABSTRACT

Both anatomical total shoulder arthroplasty (TSA) and reverse shoulder arthroplasty (RSA) are routinely performed for patients who desire to continuously work or participate in sports. This study analyzes and compares the ability of patients to work and partake in sports following shoulder arthroplasty based on responses to clinical outcome surveys.

A retrospective review of the shoulder surgery repository was performed for all patients treated with TSA and RSA and who completed questions 9 and 10 on the activity patient self-evaluation portion of the American Shoulder and Elbow Surgeons (ASES) Assessment Form. Patients with a minimum of 1-year follow-up were included if a sport or work was identified. The analysis included 162 patients with TSA and 114 patients with RSA. Comparisons were made between TSA and RSA in terms of the specific ASES scores (rated 0-3) reported for ability to work and participate in sports and total ASES scores, and scores based on specific sports or line of work reported. Comparisons were also made between sports predominantly using shoulder function and those that do not.

TSA patients had a 27% higher ability to participate in sports (average specific ASES score: 2.5 vs 1.9, P < .001) than RSA patients and presented significantly higher scores for swimming and golf. Compared with RSA patients, TSA patients demonstrated more ability to participate in sports requiring shoulder function without difficulty, as 63% reported maximal scores (P = .003). Total shoulder arthroplasty patients also demonstrated a 21% higher ability to work than RSA patients (average specific ASES scores: 2.6 vs 2.1, P < .001), yielding significantly higher scores for housework and gardening.

Both TSA and RSA allow for participation in work and sports, with TSA patients reporting better overall ability to participate. For sports involving shoulder function, TSA patients more commonly report maximal ability to participate than RSA patients.

End-stage shoulder arthritis has been successfully treated with anatomical total shoulder arthroplasty (TSA) with high rates of functional recovery.¹ With the introduction of reverse shoulder arthroplasty (RSA), indications for TSA have expanded.^2-6 With continuing expansion of surgical indications, a more diverse and potentially active patient population is now being treated. As patients exhibit increased awareness of health and wellness, they demonstrate significant interest in understanding their ability to work or participate in sports after surgery.⁷ Patients no longer focus on pain relief as the only goal of surgery. A recent study of patients aged 65 years and undergoing shoulder arthroplasty revealed that 64% of the patients listed the ability to return to sports as the main reason for undergoing surgery,⁸ highlighting the significance of sports play in a patient’s life. Prior to surgery, shoulder pathologies lead to impairment in function, range of motion, and pain,⁹ hindering a patient to participate in both work and sports. With the intervention yielding improvement to these areas^6,9-13 with increased patient satisfaction,^10,13 accurately tailoring patient expectations for participation in sports and work postoperatively becomes increasingly important.

Continue to: Although several studies...

 

Although several studies have demonstrated the ability of patients to return to sports following TSA,^8,14-18 a limited number of studies discuss the return to sports following RSA.^19-21 Despite known postoperative improvements, no clear consensus is reached as to which specific sports patients can return to and at what level of participation is to be expected. Surveyed members of the American Shoulder and Elbow Surgeons (ASES) universally favored full return to sports, except for contact sports for TSA patients, whereas other surgeons are more conservative to allow RSA patients to return to activities.²² To our knowledge, no other study has investigated the ability to work following RSA. Furthermore, no other study has used patient-reported outcomes to compare the quality of participation in sports or work between TSA and RSA patients following surgery. This study reports the ability of patients treated with TSA and RSA to work and participate in sports based on clinical outcome surveys. We hypothesize that TSA patients will be allowed to work and participate in sports with less difficulty than RSA patients.

MATERIALS AND METHODS

Following Institutional Review Board approval, a retrospective review was performed on all patients treated with TSA or RSA and who completed questions 9 and/or 10 (by score and named usual sport and/or work) on the activity patient self-evaluation portion of the ASES²³ Assessment Form between 2007 to 2014; queries were made via the Shoulder Outcomes Repository. A minimum of 12-month follow-up was required, as functional recovery has been shown to plateau or nearly plateau by 12 months.¹¹ Patients were excluded if <12 months of follow-up was available, if they failed to provide a written answer for questions 9 or 10 on the activity patient self-evaluation portion of the ASES Assessment Form, or if they required a revision shoulder arthroplasty. A single fellowship-trained shoulder and elbow surgeon performed all procedures via the same deltopectoral approach and prescribed identical postoperative rehabilitation for both TSA and RSA patients. The database query yielded 162 TSA and 114 RSA patients, for a total of 276 patients eligible for the study.

For all patients, the most recent follow-up ASES score was used. Comparisons were made between TSA and RSA for total ASES scores and response groups for usual sport (ASES question 9) and usual work (ASES question 10). The ASES questionnaire provides patients with 4 choices for each question based on the ability to perform each activity: 0, unable to do; 1, very difficult; 2, somewhat difficult; and 3, not difficult. The questionnaire also allows the patients to identify their usual work and sports. If patients noted >1 sport or work activity, they were included within multiple subgroups. Patients were further compared by age and gender.

Work was subdivided to include retired, housework, desk jobs, prolonged standing, gardening/yard work, jobs requiring lifting, carpenter/construction, cook/food preparation, and creative jobs (Table 1). 

Sports were subdivided to include golf, swimming, walking, gym exercises, racquet sports, group fitness, sedentary sports, fishing and target shooting, adventure sports, team sports, bowling, biking, hiking/jogging/rollerblading, and motorcycle riding (Table 2). 

Sports were further subdivided into those which predominately use shoulder function and those requiring minimal shoulder function (Table 3).

Statistical analysis was performed with SPSS Version 21 (IBM). Unpaired t tests were used to determine differences between groups. A P-value of <.05 was deemed significant.

Continue to: A total of 276 patients...

 

RESULTS

A total of 276 patients that met the inclusion criteria were eligible for the study, with 162 having undergone TSA and 114 with RSA. Overall average follow-up totaled 29 months (range, 12-91 months). RSA patients (average age, 75 years old; range, 46-88 years) were significantly older than TSA patients (average age, 69 years old; range, 32-89 years; P = .001). Significantly more women were treated with TSA (52% TSA; 48% RSA; P = .012), whereas significantly more men were treated with TSA (67% TSA; 33% RSA, P = .012). Total ASES scores were significantly higher for TSA patients than RSA patients in work (P = .012) (Table 4) but not in sports (P = .063) (Table 5) categories.

 

SPORTS

A total of 186 patients, comprising of 71 RSA and 115 TSA individuals, responded to question 9 of the ASES questionnaire (Table 5). Among usually reported sports, golf (25%), swimming (17%), and walking (18%) were the most commonly cited. RSA patients indicating a sport were significantly older than TSA patients (74 years vs 69 years, P < .001). TSA patients reported a 27% higher difference in overall ability to participate in sports, with an average ASES sport-specific score of 2.5 compared with the 1.9 for RSA patients (P < .001).

Among specific sports, TSA patients reported significantly higher scores for swimming (2.6 vs 1.8, P = .007) and golf (2.5 vs 1.8, P = .050). However, no significant differences were observed for walking, gym exercises, and racquet sports (Table 5). Among sport subsets, RSA patients were significantly older for golf (77 years vs 70 years, P = .006) and bowling (80 years vs 68 years, P = .005). Five TSA patients reported biking as their sport, whereas no RSA patient reported such activity. Within each subset of sports, no significant differences were noted in average ASES total scores.

TSA patients demonstrated a more significant ability to perform usual sports that involve shoulder function without difficulty (score of 3). In shoulder dominant sports, a total of 63% of TSA patients reported a score of 3 compared with the 39% of RSA patients (P = .003). RSA patients more often reported an inability to perform shoulder specific sports, as proven by 20% of RSA patients reporting a score of 0 compared with 4% of TSA patients (P < .001) (Table 6).

WORK

A total of 265 patients, including 106 RSA and 159 TSA patients, responded to question 10 of the ASES questionnaire. Among usually reported work, retirement (43%), housework (27%), and desk jobs (18%) were the most commonly cited. RSA patients denoting a work were significantly older than TSA patients (75 years vs 69 years, P < .001). Patients with TSA presented a 21% higher difference in the overall ability to work, featuring an average ASES work-specific score of 2.6 compared with the 2.1 for RSA patients (P < .001) (Table 4).

Continue to: Among specific work activities...

 

Among specific work activities, TSA patients reported significantly higher scores for housework (2.7 vs 2; 34% difference; P = .001) and gardening (2.8 vs 1.7; 65% difference; P = .009) in comparison with RSA patients. However, no significant differences were observed for other work activities, including retirement, desk job, prolonged standing, creative jobs, lifting jobs, or construction (Table 4). Among the work subgroups, RSA patients were older than TSA patients for the retired group (77 years vs 72 years; P < .001) and gardening (81 years vs 68 years; P = .002).

DISCUSSION

The ability to participate in sports and work is a common goal for shoulder arthroplasty patients. However, the ability at which participation occurs has not been examined. This study illustrates not only the ability to engage in usual work or sport, but provides some insights into patient-reported quality of participation. Overall, TSA patients featured 27% higher sport-specific ASES scores and 21% higher work-specific ASES scores than RSA patients, confirming our hypothesis that TSA patients can participate in work or sports with less difficulty in general. This study is the first to stratify the difficulty of participating in sports in general and in specific sports identified by patients. Although statistical analysis was performed for individual sports and work reported, the use of small cohorts possibly affected the ability to detect significant differences. The data presented in this study can thus be used as descriptive evidence of what a patient may expect to be able to do following surgery, helping to define patient expectations prior to electing to undergo shoulder arthroplasty.

Among specific sports identified by patients, a few significant differences were observed between RSA and TSA patients. However, ASES-specific scores almost universally favored TSA. Of the sport subgroups, swimming and golf showed significant differences. For swimming, this difference was fairly significant, as TSA patients demonstrated a 49% higher score than their RSA counterparts, but without differences in age or total ASES score (Table 5). Alteration in shoulder mechanics after RSA may be used to explain the difficulty in returning to swimming, as additional time may be needed to adapt to new mechanics.²⁴ McCarty and colleagues⁸ demonstrated that 90% of patients following TSA fully resumed participation in swimming within 6 months of surgery, and further stated that repetitive motions of swimming caused no effects on short-term outcomes. No similar analysis of swimming has been reported for RSA patients. Based upon our findings, the average RSA patient can experience some difficulties when returning to swimming after surgery (average specific ASES score, 1.8).

Jensen and Rockwood¹⁶ were among the first to demonstrate successful return to golf of 24 patients who had undergone either TSA or hemiarthroplasty (HA), showing a 5-stroke improvement in their game. A recent study investigating patient-reported activity in patients aged 75 years and undergoing RSA showed that 23% of patients returned to high-level activity sports, such as golf, motorcycle riding, or free weights.¹⁹ All patients who participated in golf before surgery resumed playing following surgery; however, golf was listed among the top activities that patients wanted to participate in but could not for any reason.¹⁹ Our data suggest that golfers with TSA will face less difficulty returning to sports compared with their RSA counterparts (average specific ASES score, 2.5 vs 1.8, who might find golf somewhat difficult.

Although no study has provided a clear consensus as to which activities are safe to perform following shoulder arthroplasty, experts have suggested that activities that impart high loads on the glenohumeral joint should be avoided.¹⁵ Among TSA patients, McCarty and colleagues⁸ reported high rates of return for swimmers, golfers, and tennis players; however, relatively low rates were reported for weight lifting, bowling, and softball (20%). Within our study group, golf, swimming, and walking were listed among the most popular sports performed. Although weight lifting, bowling, and softball were less commonly identified as usual sports within our study, patients treated with TSA demonstrated more ease to participate than RSA patients. This result was observed with ASES-specific scores noted for weight lifting and gym exercises (TSA, 2.5; RSA, 2.3) and team sports, such as softball (TSA, 2; RSA, 1.3). However, for bowling, RSA patients showed a trend toward more ability (RSA, 2.7; TSA, 1.7).

Continue to: Among specific work activities...

 

Successful return to sports that involve shoulder function, such as golf and swimming, has been demonstrated for TSA.^8,14,16,17 However, studies have reported that return to these sports can be difficult for RSA patients.²⁰ Fink and colleagues¹⁹ reported that following RSA, 48.7% of patients returned to moderate-intensity sports, such as swimming and golf. Consistent with these findings, in our study, TSA patients demonstrated a significantly higher ability to participate in their usual sports without difficulty (ASES-specific score of 3). This observation may relate to lower ultimate achievements in range of motion and strength in patients treated with RSA, when compared with TSA patients,^24,25 and the generalized practice of utilizing RSA for lower-demand patients (RSA patients in this study were older).

Overall, participation in work was 21% easier for TSA patients than RSA patients. Although the majority of our patients cited retirement as their primary work, which is consistent with what one would expect with the mean age of this study’s cohorts (RSA, 75 years; TSA, 69 years), housework and gardening were the only specifically identified forms of work that demonstrated significant differences between RSA and TSA patients. A few reports in the literature documented the ability to return to work after shoulder arthroplasty. In a recent report on 13 workers’ compensation patients treated with TSA, only 1 patient returned to the same job, and 54% did not return to work.²⁶ In a study comparing 14 workers’ compensation to a matched group of controls with all members treated with RSA, the workers’ compensation group yielded a lower return-to-work rate (14.2%) than the controls (41.7%).²⁷ In a large study of 154 TSA patients, 14% returned to work, but specific jobs were not described in this analysis.¹⁴

The results of this study suggest that more TSA patients successfully participate in low-demand activities, such as gardening or housework. Zarkadas and colleagues¹⁸ reported that 65% of TSA and 47% of HA patients successfully returned to gardening compared with 42% of RSA patients observed in a continuation study.²⁰ This study showed that TSA patients yielded a 65% difference in ability to work in gardening and 34% difference in ability to perform housework compared with RSA patients. Based on these findings, TSA patients can expect to experience no difficulty in performing housework or gardening, whereas RSA patients may find these tasks difficult to a certain degree.

The main limitation of this study is the reporting bias that results from survey-based studies. Possibly, more people engage in specific sports or work than what were reported. This type of study also features an inherent selection bias, as patients with highly and physically demanding jobs or usual sports were less likely to have been offered either TSA or RSA. An additional important limitation is the relatively small cohorts within sport and work subgroups; the small cohorts probably underpowered the statistical results of this study and made these findings valuable mostly as descriptive observations. Larger studies focusing on each subgroup will further clarify the ability of shoulder arthroplasty to perform individual sports or work. Further studies evaluating preoperative to postoperative sports- and work-specific ASES scores would provide notable insights into the functional improvements observed within each sport or work following surgery. The relatively large study population of 276 patients strengthened the findings, which relate to the overall ability to participate in sports and work for TSA and RSA patients. Finally, the evaluated TSA and RSA patients possibly represent different groups (significant difference in age and gender) with different underlying pathologies and potentially different demands and expectations. However, comparisons among these groups of patients bear importance in defining patient expectations related to surgery. Still, the ability to participate in sport or work possibly relates more to the limitations of the implant used than patient pathology. This possibility warrants further investigation.

CONCLUSION

Both TSA and RSA allow for participation in work and sports, with TSA patients reporting easier overall ability to participate. For sports involving shoulder function, TSA patients more commonly report maximal ability to participate than RSA patients.

ABSTRACT

Both anatomical total shoulder arthroplasty (TSA) and reverse shoulder arthroplasty (RSA) are routinely performed for patients who desire to continuously work or participate in sports. This study analyzes and compares the ability of patients to work and partake in sports following shoulder arthroplasty based on responses to clinical outcome surveys.

A retrospective review of the shoulder surgery repository was performed for all patients treated with TSA and RSA and who completed questions 9 and 10 on the activity patient self-evaluation portion of the American Shoulder and Elbow Surgeons (ASES) Assessment Form. Patients with a minimum of 1-year follow-up were included if a sport or work was identified. The analysis included 162 patients with TSA and 114 patients with RSA. Comparisons were made between TSA and RSA in terms of the specific ASES scores (rated 0-3) reported for ability to work and participate in sports and total ASES scores, and scores based on specific sports or line of work reported. Comparisons were also made between sports predominantly using shoulder function and those that do not.

TSA patients had a 27% higher ability to participate in sports (average specific ASES score: 2.5 vs 1.9, P < .001) than RSA patients and presented significantly higher scores for swimming and golf. Compared with RSA patients, TSA patients demonstrated more ability to participate in sports requiring shoulder function without difficulty, as 63% reported maximal scores (P = .003). Total shoulder arthroplasty patients also demonstrated a 21% higher ability to work than RSA patients (average specific ASES scores: 2.6 vs 2.1, P < .001), yielding significantly higher scores for housework and gardening.

Both TSA and RSA allow for participation in work and sports, with TSA patients reporting better overall ability to participate. For sports involving shoulder function, TSA patients more commonly report maximal ability to participate than RSA patients.

End-stage shoulder arthritis has been successfully treated with anatomical total shoulder arthroplasty (TSA) with high rates of functional recovery.¹ With the introduction of reverse shoulder arthroplasty (RSA), indications for TSA have expanded.^2-6 With continuing expansion of surgical indications, a more diverse and potentially active patient population is now being treated. As patients exhibit increased awareness of health and wellness, they demonstrate significant interest in understanding their ability to work or participate in sports after surgery.⁷ Patients no longer focus on pain relief as the only goal of surgery. A recent study of patients aged 65 years and undergoing shoulder arthroplasty revealed that 64% of the patients listed the ability to return to sports as the main reason for undergoing surgery,⁸ highlighting the significance of sports play in a patient’s life. Prior to surgery, shoulder pathologies lead to impairment in function, range of motion, and pain,⁹ hindering a patient to participate in both work and sports. With the intervention yielding improvement to these areas^6,9-13 with increased patient satisfaction,^10,13 accurately tailoring patient expectations for participation in sports and work postoperatively becomes increasingly important.

Continue to: Although several studies...

 

Although several studies have demonstrated the ability of patients to return to sports following TSA,^8,14-18 a limited number of studies discuss the return to sports following RSA.^19-21 Despite known postoperative improvements, no clear consensus is reached as to which specific sports patients can return to and at what level of participation is to be expected. Surveyed members of the American Shoulder and Elbow Surgeons (ASES) universally favored full return to sports, except for contact sports for TSA patients, whereas other surgeons are more conservative to allow RSA patients to return to activities.²² To our knowledge, no other study has investigated the ability to work following RSA. Furthermore, no other study has used patient-reported outcomes to compare the quality of participation in sports or work between TSA and RSA patients following surgery. This study reports the ability of patients treated with TSA and RSA to work and participate in sports based on clinical outcome surveys. We hypothesize that TSA patients will be allowed to work and participate in sports with less difficulty than RSA patients.

MATERIALS AND METHODS

Following Institutional Review Board approval, a retrospective review was performed on all patients treated with TSA or RSA and who completed questions 9 and/or 10 (by score and named usual sport and/or work) on the activity patient self-evaluation portion of the ASES²³ Assessment Form between 2007 to 2014; queries were made via the Shoulder Outcomes Repository. A minimum of 12-month follow-up was required, as functional recovery has been shown to plateau or nearly plateau by 12 months.¹¹ Patients were excluded if <12 months of follow-up was available, if they failed to provide a written answer for questions 9 or 10 on the activity patient self-evaluation portion of the ASES Assessment Form, or if they required a revision shoulder arthroplasty. A single fellowship-trained shoulder and elbow surgeon performed all procedures via the same deltopectoral approach and prescribed identical postoperative rehabilitation for both TSA and RSA patients. The database query yielded 162 TSA and 114 RSA patients, for a total of 276 patients eligible for the study.

For all patients, the most recent follow-up ASES score was used. Comparisons were made between TSA and RSA for total ASES scores and response groups for usual sport (ASES question 9) and usual work (ASES question 10). The ASES questionnaire provides patients with 4 choices for each question based on the ability to perform each activity: 0, unable to do; 1, very difficult; 2, somewhat difficult; and 3, not difficult. The questionnaire also allows the patients to identify their usual work and sports. If patients noted >1 sport or work activity, they were included within multiple subgroups. Patients were further compared by age and gender.

Work was subdivided to include retired, housework, desk jobs, prolonged standing, gardening/yard work, jobs requiring lifting, carpenter/construction, cook/food preparation, and creative jobs (Table 1). 

Sports were subdivided to include golf, swimming, walking, gym exercises, racquet sports, group fitness, sedentary sports, fishing and target shooting, adventure sports, team sports, bowling, biking, hiking/jogging/rollerblading, and motorcycle riding (Table 2). 

Sports were further subdivided into those which predominately use shoulder function and those requiring minimal shoulder function (Table 3).

Statistical analysis was performed with SPSS Version 21 (IBM). Unpaired t tests were used to determine differences between groups. A P-value of <.05 was deemed significant.

Continue to: A total of 276 patients...

 

RESULTS

A total of 276 patients that met the inclusion criteria were eligible for the study, with 162 having undergone TSA and 114 with RSA. Overall average follow-up totaled 29 months (range, 12-91 months). RSA patients (average age, 75 years old; range, 46-88 years) were significantly older than TSA patients (average age, 69 years old; range, 32-89 years; P = .001). Significantly more women were treated with TSA (52% TSA; 48% RSA; P = .012), whereas significantly more men were treated with TSA (67% TSA; 33% RSA, P = .012). Total ASES scores were significantly higher for TSA patients than RSA patients in work (P = .012) (Table 4) but not in sports (P = .063) (Table 5) categories.

 

SPORTS

A total of 186 patients, comprising of 71 RSA and 115 TSA individuals, responded to question 9 of the ASES questionnaire (Table 5). Among usually reported sports, golf (25%), swimming (17%), and walking (18%) were the most commonly cited. RSA patients indicating a sport were significantly older than TSA patients (74 years vs 69 years, P < .001). TSA patients reported a 27% higher difference in overall ability to participate in sports, with an average ASES sport-specific score of 2.5 compared with the 1.9 for RSA patients (P < .001).

Among specific sports, TSA patients reported significantly higher scores for swimming (2.6 vs 1.8, P = .007) and golf (2.5 vs 1.8, P = .050). However, no significant differences were observed for walking, gym exercises, and racquet sports (Table 5). Among sport subsets, RSA patients were significantly older for golf (77 years vs 70 years, P = .006) and bowling (80 years vs 68 years, P = .005). Five TSA patients reported biking as their sport, whereas no RSA patient reported such activity. Within each subset of sports, no significant differences were noted in average ASES total scores.

TSA patients demonstrated a more significant ability to perform usual sports that involve shoulder function without difficulty (score of 3). In shoulder dominant sports, a total of 63% of TSA patients reported a score of 3 compared with the 39% of RSA patients (P = .003). RSA patients more often reported an inability to perform shoulder specific sports, as proven by 20% of RSA patients reporting a score of 0 compared with 4% of TSA patients (P < .001) (Table 6).

WORK

A total of 265 patients, including 106 RSA and 159 TSA patients, responded to question 10 of the ASES questionnaire. Among usually reported work, retirement (43%), housework (27%), and desk jobs (18%) were the most commonly cited. RSA patients denoting a work were significantly older than TSA patients (75 years vs 69 years, P < .001). Patients with TSA presented a 21% higher difference in the overall ability to work, featuring an average ASES work-specific score of 2.6 compared with the 2.1 for RSA patients (P < .001) (Table 4).

Continue to: Among specific work activities...

 

Among specific work activities, TSA patients reported significantly higher scores for housework (2.7 vs 2; 34% difference; P = .001) and gardening (2.8 vs 1.7; 65% difference; P = .009) in comparison with RSA patients. However, no significant differences were observed for other work activities, including retirement, desk job, prolonged standing, creative jobs, lifting jobs, or construction (Table 4). Among the work subgroups, RSA patients were older than TSA patients for the retired group (77 years vs 72 years; P < .001) and gardening (81 years vs 68 years; P = .002).

DISCUSSION

The ability to participate in sports and work is a common goal for shoulder arthroplasty patients. However, the ability at which participation occurs has not been examined. This study illustrates not only the ability to engage in usual work or sport, but provides some insights into patient-reported quality of participation. Overall, TSA patients featured 27% higher sport-specific ASES scores and 21% higher work-specific ASES scores than RSA patients, confirming our hypothesis that TSA patients can participate in work or sports with less difficulty in general. This study is the first to stratify the difficulty of participating in sports in general and in specific sports identified by patients. Although statistical analysis was performed for individual sports and work reported, the use of small cohorts possibly affected the ability to detect significant differences. The data presented in this study can thus be used as descriptive evidence of what a patient may expect to be able to do following surgery, helping to define patient expectations prior to electing to undergo shoulder arthroplasty.

Among specific sports identified by patients, a few significant differences were observed between RSA and TSA patients. However, ASES-specific scores almost universally favored TSA. Of the sport subgroups, swimming and golf showed significant differences. For swimming, this difference was fairly significant, as TSA patients demonstrated a 49% higher score than their RSA counterparts, but without differences in age or total ASES score (Table 5). Alteration in shoulder mechanics after RSA may be used to explain the difficulty in returning to swimming, as additional time may be needed to adapt to new mechanics.²⁴ McCarty and colleagues⁸ demonstrated that 90% of patients following TSA fully resumed participation in swimming within 6 months of surgery, and further stated that repetitive motions of swimming caused no effects on short-term outcomes. No similar analysis of swimming has been reported for RSA patients. Based upon our findings, the average RSA patient can experience some difficulties when returning to swimming after surgery (average specific ASES score, 1.8).

Jensen and Rockwood¹⁶ were among the first to demonstrate successful return to golf of 24 patients who had undergone either TSA or hemiarthroplasty (HA), showing a 5-stroke improvement in their game. A recent study investigating patient-reported activity in patients aged 75 years and undergoing RSA showed that 23% of patients returned to high-level activity sports, such as golf, motorcycle riding, or free weights.¹⁹ All patients who participated in golf before surgery resumed playing following surgery; however, golf was listed among the top activities that patients wanted to participate in but could not for any reason.¹⁹ Our data suggest that golfers with TSA will face less difficulty returning to sports compared with their RSA counterparts (average specific ASES score, 2.5 vs 1.8, who might find golf somewhat difficult.

Although no study has provided a clear consensus as to which activities are safe to perform following shoulder arthroplasty, experts have suggested that activities that impart high loads on the glenohumeral joint should be avoided.¹⁵ Among TSA patients, McCarty and colleagues⁸ reported high rates of return for swimmers, golfers, and tennis players; however, relatively low rates were reported for weight lifting, bowling, and softball (20%). Within our study group, golf, swimming, and walking were listed among the most popular sports performed. Although weight lifting, bowling, and softball were less commonly identified as usual sports within our study, patients treated with TSA demonstrated more ease to participate than RSA patients. This result was observed with ASES-specific scores noted for weight lifting and gym exercises (TSA, 2.5; RSA, 2.3) and team sports, such as softball (TSA, 2; RSA, 1.3). However, for bowling, RSA patients showed a trend toward more ability (RSA, 2.7; TSA, 1.7).

Continue to: Among specific work activities...

 

Successful return to sports that involve shoulder function, such as golf and swimming, has been demonstrated for TSA.^8,14,16,17 However, studies have reported that return to these sports can be difficult for RSA patients.²⁰ Fink and colleagues¹⁹ reported that following RSA, 48.7% of patients returned to moderate-intensity sports, such as swimming and golf. Consistent with these findings, in our study, TSA patients demonstrated a significantly higher ability to participate in their usual sports without difficulty (ASES-specific score of 3). This observation may relate to lower ultimate achievements in range of motion and strength in patients treated with RSA, when compared with TSA patients,^24,25 and the generalized practice of utilizing RSA for lower-demand patients (RSA patients in this study were older).

Overall, participation in work was 21% easier for TSA patients than RSA patients. Although the majority of our patients cited retirement as their primary work, which is consistent with what one would expect with the mean age of this study’s cohorts (RSA, 75 years; TSA, 69 years), housework and gardening were the only specifically identified forms of work that demonstrated significant differences between RSA and TSA patients. A few reports in the literature documented the ability to return to work after shoulder arthroplasty. In a recent report on 13 workers’ compensation patients treated with TSA, only 1 patient returned to the same job, and 54% did not return to work.²⁶ In a study comparing 14 workers’ compensation to a matched group of controls with all members treated with RSA, the workers’ compensation group yielded a lower return-to-work rate (14.2%) than the controls (41.7%).²⁷ In a large study of 154 TSA patients, 14% returned to work, but specific jobs were not described in this analysis.¹⁴

The results of this study suggest that more TSA patients successfully participate in low-demand activities, such as gardening or housework. Zarkadas and colleagues¹⁸ reported that 65% of TSA and 47% of HA patients successfully returned to gardening compared with 42% of RSA patients observed in a continuation study.²⁰ This study showed that TSA patients yielded a 65% difference in ability to work in gardening and 34% difference in ability to perform housework compared with RSA patients. Based on these findings, TSA patients can expect to experience no difficulty in performing housework or gardening, whereas RSA patients may find these tasks difficult to a certain degree.

The main limitation of this study is the reporting bias that results from survey-based studies. Possibly, more people engage in specific sports or work than what were reported. This type of study also features an inherent selection bias, as patients with highly and physically demanding jobs or usual sports were less likely to have been offered either TSA or RSA. An additional important limitation is the relatively small cohorts within sport and work subgroups; the small cohorts probably underpowered the statistical results of this study and made these findings valuable mostly as descriptive observations. Larger studies focusing on each subgroup will further clarify the ability of shoulder arthroplasty to perform individual sports or work. Further studies evaluating preoperative to postoperative sports- and work-specific ASES scores would provide notable insights into the functional improvements observed within each sport or work following surgery. The relatively large study population of 276 patients strengthened the findings, which relate to the overall ability to participate in sports and work for TSA and RSA patients. Finally, the evaluated TSA and RSA patients possibly represent different groups (significant difference in age and gender) with different underlying pathologies and potentially different demands and expectations. However, comparisons among these groups of patients bear importance in defining patient expectations related to surgery. Still, the ability to participate in sport or work possibly relates more to the limitations of the implant used than patient pathology. This possibility warrants further investigation.

CONCLUSION

Both TSA and RSA allow for participation in work and sports, with TSA patients reporting easier overall ability to participate. For sports involving shoulder function, TSA patients more commonly report maximal ability to participate than RSA patients.

ABSTRACT

Both anatomical total shoulder arthroplasty (TSA) and reverse shoulder arthroplasty (RSA) are routinely performed for patients who desire to continuously work or participate in sports. This study analyzes and compares the ability of patients to work and partake in sports following shoulder arthroplasty based on responses to clinical outcome surveys.

A retrospective review of the shoulder surgery repository was performed for all patients treated with TSA and RSA and who completed questions 9 and 10 on the activity patient self-evaluation portion of the American Shoulder and Elbow Surgeons (ASES) Assessment Form. Patients with a minimum of 1-year follow-up were included if a sport or work was identified. The analysis included 162 patients with TSA and 114 patients with RSA. Comparisons were made between TSA and RSA in terms of the specific ASES scores (rated 0-3) reported for ability to work and participate in sports and total ASES scores, and scores based on specific sports or line of work reported. Comparisons were also made between sports predominantly using shoulder function and those that do not.

TSA patients had a 27% higher ability to participate in sports (average specific ASES score: 2.5 vs 1.9, P < .001) than RSA patients and presented significantly higher scores for swimming and golf. Compared with RSA patients, TSA patients demonstrated more ability to participate in sports requiring shoulder function without difficulty, as 63% reported maximal scores (P = .003). Total shoulder arthroplasty patients also demonstrated a 21% higher ability to work than RSA patients (average specific ASES scores: 2.6 vs 2.1, P < .001), yielding significantly higher scores for housework and gardening.

Both TSA and RSA allow for participation in work and sports, with TSA patients reporting better overall ability to participate. For sports involving shoulder function, TSA patients more commonly report maximal ability to participate than RSA patients.

End-stage shoulder arthritis has been successfully treated with anatomical total shoulder arthroplasty (TSA) with high rates of functional recovery.¹ With the introduction of reverse shoulder arthroplasty (RSA), indications for TSA have expanded.^2-6 With continuing expansion of surgical indications, a more diverse and potentially active patient population is now being treated. As patients exhibit increased awareness of health and wellness, they demonstrate significant interest in understanding their ability to work or participate in sports after surgery.⁷ Patients no longer focus on pain relief as the only goal of surgery. A recent study of patients aged 65 years and undergoing shoulder arthroplasty revealed that 64% of the patients listed the ability to return to sports as the main reason for undergoing surgery,⁸ highlighting the significance of sports play in a patient’s life. Prior to surgery, shoulder pathologies lead to impairment in function, range of motion, and pain,⁹ hindering a patient to participate in both work and sports. With the intervention yielding improvement to these areas^6,9-13 with increased patient satisfaction,^10,13 accurately tailoring patient expectations for participation in sports and work postoperatively becomes increasingly important.

Continue to: Although several studies...

 

Although several studies have demonstrated the ability of patients to return to sports following TSA,^8,14-18 a limited number of studies discuss the return to sports following RSA.^19-21 Despite known postoperative improvements, no clear consensus is reached as to which specific sports patients can return to and at what level of participation is to be expected. Surveyed members of the American Shoulder and Elbow Surgeons (ASES) universally favored full return to sports, except for contact sports for TSA patients, whereas other surgeons are more conservative to allow RSA patients to return to activities.²² To our knowledge, no other study has investigated the ability to work following RSA. Furthermore, no other study has used patient-reported outcomes to compare the quality of participation in sports or work between TSA and RSA patients following surgery. This study reports the ability of patients treated with TSA and RSA to work and participate in sports based on clinical outcome surveys. We hypothesize that TSA patients will be allowed to work and participate in sports with less difficulty than RSA patients.

MATERIALS AND METHODS

Following Institutional Review Board approval, a retrospective review was performed on all patients treated with TSA or RSA and who completed questions 9 and/or 10 (by score and named usual sport and/or work) on the activity patient self-evaluation portion of the ASES²³ Assessment Form between 2007 to 2014; queries were made via the Shoulder Outcomes Repository. A minimum of 12-month follow-up was required, as functional recovery has been shown to plateau or nearly plateau by 12 months.¹¹ Patients were excluded if <12 months of follow-up was available, if they failed to provide a written answer for questions 9 or 10 on the activity patient self-evaluation portion of the ASES Assessment Form, or if they required a revision shoulder arthroplasty. A single fellowship-trained shoulder and elbow surgeon performed all procedures via the same deltopectoral approach and prescribed identical postoperative rehabilitation for both TSA and RSA patients. The database query yielded 162 TSA and 114 RSA patients, for a total of 276 patients eligible for the study.

For all patients, the most recent follow-up ASES score was used. Comparisons were made between TSA and RSA for total ASES scores and response groups for usual sport (ASES question 9) and usual work (ASES question 10). The ASES questionnaire provides patients with 4 choices for each question based on the ability to perform each activity: 0, unable to do; 1, very difficult; 2, somewhat difficult; and 3, not difficult. The questionnaire also allows the patients to identify their usual work and sports. If patients noted >1 sport or work activity, they were included within multiple subgroups. Patients were further compared by age and gender.

Work was subdivided to include retired, housework, desk jobs, prolonged standing, gardening/yard work, jobs requiring lifting, carpenter/construction, cook/food preparation, and creative jobs (Table 1). 

Sports were subdivided to include golf, swimming, walking, gym exercises, racquet sports, group fitness, sedentary sports, fishing and target shooting, adventure sports, team sports, bowling, biking, hiking/jogging/rollerblading, and motorcycle riding (Table 2). 

Sports were further subdivided into those which predominately use shoulder function and those requiring minimal shoulder function (Table 3).

Statistical analysis was performed with SPSS Version 21 (IBM). Unpaired t tests were used to determine differences between groups. A P-value of <.05 was deemed significant.

Continue to: A total of 276 patients...

 

RESULTS

A total of 276 patients that met the inclusion criteria were eligible for the study, with 162 having undergone TSA and 114 with RSA. Overall average follow-up totaled 29 months (range, 12-91 months). RSA patients (average age, 75 years old; range, 46-88 years) were significantly older than TSA patients (average age, 69 years old; range, 32-89 years; P = .001). Significantly more women were treated with TSA (52% TSA; 48% RSA; P = .012), whereas significantly more men were treated with TSA (67% TSA; 33% RSA, P = .012). Total ASES scores were significantly higher for TSA patients than RSA patients in work (P = .012) (Table 4) but not in sports (P = .063) (Table 5) categories.

 

SPORTS

A total of 186 patients, comprising of 71 RSA and 115 TSA individuals, responded to question 9 of the ASES questionnaire (Table 5). Among usually reported sports, golf (25%), swimming (17%), and walking (18%) were the most commonly cited. RSA patients indicating a sport were significantly older than TSA patients (74 years vs 69 years, P < .001). TSA patients reported a 27% higher difference in overall ability to participate in sports, with an average ASES sport-specific score of 2.5 compared with the 1.9 for RSA patients (P < .001).

Among specific sports, TSA patients reported significantly higher scores for swimming (2.6 vs 1.8, P = .007) and golf (2.5 vs 1.8, P = .050). However, no significant differences were observed for walking, gym exercises, and racquet sports (Table 5). Among sport subsets, RSA patients were significantly older for golf (77 years vs 70 years, P = .006) and bowling (80 years vs 68 years, P = .005). Five TSA patients reported biking as their sport, whereas no RSA patient reported such activity. Within each subset of sports, no significant differences were noted in average ASES total scores.

TSA patients demonstrated a more significant ability to perform usual sports that involve shoulder function without difficulty (score of 3). In shoulder dominant sports, a total of 63% of TSA patients reported a score of 3 compared with the 39% of RSA patients (P = .003). RSA patients more often reported an inability to perform shoulder specific sports, as proven by 20% of RSA patients reporting a score of 0 compared with 4% of TSA patients (P < .001) (Table 6).

WORK

A total of 265 patients, including 106 RSA and 159 TSA patients, responded to question 10 of the ASES questionnaire. Among usually reported work, retirement (43%), housework (27%), and desk jobs (18%) were the most commonly cited. RSA patients denoting a work were significantly older than TSA patients (75 years vs 69 years, P < .001). Patients with TSA presented a 21% higher difference in the overall ability to work, featuring an average ASES work-specific score of 2.6 compared with the 2.1 for RSA patients (P < .001) (Table 4).

Continue to: Among specific work activities...

 

Among specific work activities, TSA patients reported significantly higher scores for housework (2.7 vs 2; 34% difference; P = .001) and gardening (2.8 vs 1.7; 65% difference; P = .009) in comparison with RSA patients. However, no significant differences were observed for other work activities, including retirement, desk job, prolonged standing, creative jobs, lifting jobs, or construction (Table 4). Among the work subgroups, RSA patients were older than TSA patients for the retired group (77 years vs 72 years; P < .001) and gardening (81 years vs 68 years; P = .002).

DISCUSSION

The ability to participate in sports and work is a common goal for shoulder arthroplasty patients. However, the ability at which participation occurs has not been examined. This study illustrates not only the ability to engage in usual work or sport, but provides some insights into patient-reported quality of participation. Overall, TSA patients featured 27% higher sport-specific ASES scores and 21% higher work-specific ASES scores than RSA patients, confirming our hypothesis that TSA patients can participate in work or sports with less difficulty in general. This study is the first to stratify the difficulty of participating in sports in general and in specific sports identified by patients. Although statistical analysis was performed for individual sports and work reported, the use of small cohorts possibly affected the ability to detect significant differences. The data presented in this study can thus be used as descriptive evidence of what a patient may expect to be able to do following surgery, helping to define patient expectations prior to electing to undergo shoulder arthroplasty.

Among specific sports identified by patients, a few significant differences were observed between RSA and TSA patients. However, ASES-specific scores almost universally favored TSA. Of the sport subgroups, swimming and golf showed significant differences. For swimming, this difference was fairly significant, as TSA patients demonstrated a 49% higher score than their RSA counterparts, but without differences in age or total ASES score (Table 5). Alteration in shoulder mechanics after RSA may be used to explain the difficulty in returning to swimming, as additional time may be needed to adapt to new mechanics.²⁴ McCarty and colleagues⁸ demonstrated that 90% of patients following TSA fully resumed participation in swimming within 6 months of surgery, and further stated that repetitive motions of swimming caused no effects on short-term outcomes. No similar analysis of swimming has been reported for RSA patients. Based upon our findings, the average RSA patient can experience some difficulties when returning to swimming after surgery (average specific ASES score, 1.8).

Jensen and Rockwood¹⁶ were among the first to demonstrate successful return to golf of 24 patients who had undergone either TSA or hemiarthroplasty (HA), showing a 5-stroke improvement in their game. A recent study investigating patient-reported activity in patients aged 75 years and undergoing RSA showed that 23% of patients returned to high-level activity sports, such as golf, motorcycle riding, or free weights.¹⁹ All patients who participated in golf before surgery resumed playing following surgery; however, golf was listed among the top activities that patients wanted to participate in but could not for any reason.¹⁹ Our data suggest that golfers with TSA will face less difficulty returning to sports compared with their RSA counterparts (average specific ASES score, 2.5 vs 1.8, who might find golf somewhat difficult.

Although no study has provided a clear consensus as to which activities are safe to perform following shoulder arthroplasty, experts have suggested that activities that impart high loads on the glenohumeral joint should be avoided.¹⁵ Among TSA patients, McCarty and colleagues⁸ reported high rates of return for swimmers, golfers, and tennis players; however, relatively low rates were reported for weight lifting, bowling, and softball (20%). Within our study group, golf, swimming, and walking were listed among the most popular sports performed. Although weight lifting, bowling, and softball were less commonly identified as usual sports within our study, patients treated with TSA demonstrated more ease to participate than RSA patients. This result was observed with ASES-specific scores noted for weight lifting and gym exercises (TSA, 2.5; RSA, 2.3) and team sports, such as softball (TSA, 2; RSA, 1.3). However, for bowling, RSA patients showed a trend toward more ability (RSA, 2.7; TSA, 1.7).

Continue to: Among specific work activities...

 

Successful return to sports that involve shoulder function, such as golf and swimming, has been demonstrated for TSA.^8,14,16,17 However, studies have reported that return to these sports can be difficult for RSA patients.²⁰ Fink and colleagues¹⁹ reported that following RSA, 48.7% of patients returned to moderate-intensity sports, such as swimming and golf. Consistent with these findings, in our study, TSA patients demonstrated a significantly higher ability to participate in their usual sports without difficulty (ASES-specific score of 3). This observation may relate to lower ultimate achievements in range of motion and strength in patients treated with RSA, when compared with TSA patients,^24,25 and the generalized practice of utilizing RSA for lower-demand patients (RSA patients in this study were older).

Overall, participation in work was 21% easier for TSA patients than RSA patients. Although the majority of our patients cited retirement as their primary work, which is consistent with what one would expect with the mean age of this study’s cohorts (RSA, 75 years; TSA, 69 years), housework and gardening were the only specifically identified forms of work that demonstrated significant differences between RSA and TSA patients. A few reports in the literature documented the ability to return to work after shoulder arthroplasty. In a recent report on 13 workers’ compensation patients treated with TSA, only 1 patient returned to the same job, and 54% did not return to work.²⁶ In a study comparing 14 workers’ compensation to a matched group of controls with all members treated with RSA, the workers’ compensation group yielded a lower return-to-work rate (14.2%) than the controls (41.7%).²⁷ In a large study of 154 TSA patients, 14% returned to work, but specific jobs were not described in this analysis.¹⁴

The results of this study suggest that more TSA patients successfully participate in low-demand activities, such as gardening or housework. Zarkadas and colleagues¹⁸ reported that 65% of TSA and 47% of HA patients successfully returned to gardening compared with 42% of RSA patients observed in a continuation study.²⁰ This study showed that TSA patients yielded a 65% difference in ability to work in gardening and 34% difference in ability to perform housework compared with RSA patients. Based on these findings, TSA patients can expect to experience no difficulty in performing housework or gardening, whereas RSA patients may find these tasks difficult to a certain degree.

The main limitation of this study is the reporting bias that results from survey-based studies. Possibly, more people engage in specific sports or work than what were reported. This type of study also features an inherent selection bias, as patients with highly and physically demanding jobs or usual sports were less likely to have been offered either TSA or RSA. An additional important limitation is the relatively small cohorts within sport and work subgroups; the small cohorts probably underpowered the statistical results of this study and made these findings valuable mostly as descriptive observations. Larger studies focusing on each subgroup will further clarify the ability of shoulder arthroplasty to perform individual sports or work. Further studies evaluating preoperative to postoperative sports- and work-specific ASES scores would provide notable insights into the functional improvements observed within each sport or work following surgery. The relatively large study population of 276 patients strengthened the findings, which relate to the overall ability to participate in sports and work for TSA and RSA patients. Finally, the evaluated TSA and RSA patients possibly represent different groups (significant difference in age and gender) with different underlying pathologies and potentially different demands and expectations. However, comparisons among these groups of patients bear importance in defining patient expectations related to surgery. Still, the ability to participate in sport or work possibly relates more to the limitations of the implant used than patient pathology. This possibility warrants further investigation.

CONCLUSION

Both TSA and RSA allow for participation in work and sports, with TSA patients reporting easier overall ability to participate. For sports involving shoulder function, TSA patients more commonly report maximal ability to participate than RSA patients.

ABSTRACT

Injuries to the distal biceps occur at the tendinous insertion at the radial tuberosity. Distal biceps injuries range from tendinosis to partial tears to non-retracted and retracted complete tears. Acute and chronic complete tears result from a tendinous avulsion at the radial tuberosity. Acute tears result from a strong force exerted on an eccentric biceps contraction, leading to tendon injury.

Distal biceps tendon injuries are uncommon (1.2 per 100,000 patients in one study).¹ An underlying degenerative component is involved in all distal biceps tendon tears and tendinosis.² Partial tears can be caused by the same mechanism or by no particular inciting event.³ Magnetic resonance imaging (MRI) is the optimal imaging modality for distal tendon tears because of its excellent specificity and sensitivity in the detection of complete tears.^4,5 Imaging also accurately diagnoses and characterizes partial tears and tendinosis.⁵ On MRI, fast spin-echo intermediate-weighted and T2-weighted or short tau inversion recovery (STIR) sequences are normally obtained to assess tendon integrity. Along with standard axial and sagittal views, the FABS (flexed elbow, abducted shoulder, supinated forearm) view is an important tool in the diagnosis of distal biceps tendon tears.⁶ The FABS view is obtained with the patient prone with the shoulder abducted 180° (above the head), with the elbow flexed to 90°, and the forearm supinated. This position allows a longitudinal view of along the entire length of the distal tendon.

Complete distal biceps tears can usually be diagnosed by history and physical examinations. However, imaging can be helpful when intact brachialis function can compensate for a completely torn tendon. MRI is also useful in the setting of a complete tear to locate the torn tendon stump, and assess the degree of retraction for tendon retrieval^7,8 and quality of the tendon stump for repair. For associated rupture of the lacertus, the degree of proximal tendon retraction can be significant (Figures 1A, 1B). 

Given that distal biceps tendon rupture occurs as an avulsion at the tendon-bone interface (Figure 2), complete distal biceps tendon tears typically demonstrate no tendon at the insertion on the radial tuberosity with a fluid-filled tendon gap with edema and/or hemorrhage^7,9 or an ill-defined T2-hyperintense mass at the expected site of the tendon.⁷ 

Complete tears without rupture of the lacertus fibrosis (bicipital aponeurosis) will have a small amount of retraction because the intact aponeurosis tethers the torn tendon stump (Figures 3A-3C). 

Chronic complete tears demonstrate heterogeneous signal intensity and fluid signal at the tendon, as well as muscle belly atrophy.⁹ A small percentage of distal biceps brachii tendons are bifid ¹⁰ (Figure 4). 

When injured, 75% have complete rupture of the short head with 17% of these having additional complete rupture of the long head, whereas 50% of those with complete rupture of the short head have partial tear or tendinosis of the long head.

Continue to: Partial distal bicep tears...

 

Partial distal bicep tears are characterized on MRI by focal or partial detachment of the tendon at the radial tuberosity with fluid filling the site of the tear. The degree of partial tearing can be assessed on MRI (Figures 5A, 5B). 

In distal biceps tendinosis, increased signals of thickened tendon fibers at the radial tuberosity are evident without focal discontinuity^7,8 (Figures 6A-6C). Patients may display attenuation of the distal tendon fibers or adjacent fluid distension representing bicipitoradial bursitis (Figures 7A, 7B).

MRI is useful in assessing the distal biceps tendon in the postoperative setting to evaluate the integrity of a repaired tendon. Cortical fixation button technique for repair creates minimal susceptibility artifacts on MRI. Postoperative MRI typically demonstrates a transverse hole drilled through the proximal radius at the site of the tuberosity with a cortical fixation button flush against the posterior radial cortex (Figures 8A-8D). 

The postoperative complication of heterotopic ossification can occasionally be observed on plain radiograph at the site of surgery, but it is less common with the current surgical technique than in the past.¹¹

ABSTRACT

Injuries to the distal biceps occur at the tendinous insertion at the radial tuberosity. Distal biceps injuries range from tendinosis to partial tears to non-retracted and retracted complete tears. Acute and chronic complete tears result from a tendinous avulsion at the radial tuberosity. Acute tears result from a strong force exerted on an eccentric biceps contraction, leading to tendon injury.

Distal biceps tendon injuries are uncommon (1.2 per 100,000 patients in one study).¹ An underlying degenerative component is involved in all distal biceps tendon tears and tendinosis.² Partial tears can be caused by the same mechanism or by no particular inciting event.³ Magnetic resonance imaging (MRI) is the optimal imaging modality for distal tendon tears because of its excellent specificity and sensitivity in the detection of complete tears.^4,5 Imaging also accurately diagnoses and characterizes partial tears and tendinosis.⁵ On MRI, fast spin-echo intermediate-weighted and T2-weighted or short tau inversion recovery (STIR) sequences are normally obtained to assess tendon integrity. Along with standard axial and sagittal views, the FABS (flexed elbow, abducted shoulder, supinated forearm) view is an important tool in the diagnosis of distal biceps tendon tears.⁶ The FABS view is obtained with the patient prone with the shoulder abducted 180° (above the head), with the elbow flexed to 90°, and the forearm supinated. This position allows a longitudinal view of along the entire length of the distal tendon.

Complete distal biceps tears can usually be diagnosed by history and physical examinations. However, imaging can be helpful when intact brachialis function can compensate for a completely torn tendon. MRI is also useful in the setting of a complete tear to locate the torn tendon stump, and assess the degree of retraction for tendon retrieval^7,8 and quality of the tendon stump for repair. For associated rupture of the lacertus, the degree of proximal tendon retraction can be significant (Figures 1A, 1B). 

Given that distal biceps tendon rupture occurs as an avulsion at the tendon-bone interface (Figure 2), complete distal biceps tendon tears typically demonstrate no tendon at the insertion on the radial tuberosity with a fluid-filled tendon gap with edema and/or hemorrhage^7,9 or an ill-defined T2-hyperintense mass at the expected site of the tendon.⁷ 

Complete tears without rupture of the lacertus fibrosis (bicipital aponeurosis) will have a small amount of retraction because the intact aponeurosis tethers the torn tendon stump (Figures 3A-3C). 

Chronic complete tears demonstrate heterogeneous signal intensity and fluid signal at the tendon, as well as muscle belly atrophy.⁹ A small percentage of distal biceps brachii tendons are bifid ¹⁰ (Figure 4). 

When injured, 75% have complete rupture of the short head with 17% of these having additional complete rupture of the long head, whereas 50% of those with complete rupture of the short head have partial tear or tendinosis of the long head.

Continue to: Partial distal bicep tears...

 

Partial distal bicep tears are characterized on MRI by focal or partial detachment of the tendon at the radial tuberosity with fluid filling the site of the tear. The degree of partial tearing can be assessed on MRI (Figures 5A, 5B). 

In distal biceps tendinosis, increased signals of thickened tendon fibers at the radial tuberosity are evident without focal discontinuity^7,8 (Figures 6A-6C). Patients may display attenuation of the distal tendon fibers or adjacent fluid distension representing bicipitoradial bursitis (Figures 7A, 7B).

MRI is useful in assessing the distal biceps tendon in the postoperative setting to evaluate the integrity of a repaired tendon. Cortical fixation button technique for repair creates minimal susceptibility artifacts on MRI. Postoperative MRI typically demonstrates a transverse hole drilled through the proximal radius at the site of the tuberosity with a cortical fixation button flush against the posterior radial cortex (Figures 8A-8D). 

The postoperative complication of heterotopic ossification can occasionally be observed on plain radiograph at the site of surgery, but it is less common with the current surgical technique than in the past.¹¹

ABSTRACT

Injuries to the distal biceps occur at the tendinous insertion at the radial tuberosity. Distal biceps injuries range from tendinosis to partial tears to non-retracted and retracted complete tears. Acute and chronic complete tears result from a tendinous avulsion at the radial tuberosity. Acute tears result from a strong force exerted on an eccentric biceps contraction, leading to tendon injury.

Distal biceps tendon injuries are uncommon (1.2 per 100,000 patients in one study).¹ An underlying degenerative component is involved in all distal biceps tendon tears and tendinosis.² Partial tears can be caused by the same mechanism or by no particular inciting event.³ Magnetic resonance imaging (MRI) is the optimal imaging modality for distal tendon tears because of its excellent specificity and sensitivity in the detection of complete tears.^4,5 Imaging also accurately diagnoses and characterizes partial tears and tendinosis.⁵ On MRI, fast spin-echo intermediate-weighted and T2-weighted or short tau inversion recovery (STIR) sequences are normally obtained to assess tendon integrity. Along with standard axial and sagittal views, the FABS (flexed elbow, abducted shoulder, supinated forearm) view is an important tool in the diagnosis of distal biceps tendon tears.⁶ The FABS view is obtained with the patient prone with the shoulder abducted 180° (above the head), with the elbow flexed to 90°, and the forearm supinated. This position allows a longitudinal view of along the entire length of the distal tendon.

Complete distal biceps tears can usually be diagnosed by history and physical examinations. However, imaging can be helpful when intact brachialis function can compensate for a completely torn tendon. MRI is also useful in the setting of a complete tear to locate the torn tendon stump, and assess the degree of retraction for tendon retrieval^7,8 and quality of the tendon stump for repair. For associated rupture of the lacertus, the degree of proximal tendon retraction can be significant (Figures 1A, 1B). 

Given that distal biceps tendon rupture occurs as an avulsion at the tendon-bone interface (Figure 2), complete distal biceps tendon tears typically demonstrate no tendon at the insertion on the radial tuberosity with a fluid-filled tendon gap with edema and/or hemorrhage^7,9 or an ill-defined T2-hyperintense mass at the expected site of the tendon.⁷ 

Complete tears without rupture of the lacertus fibrosis (bicipital aponeurosis) will have a small amount of retraction because the intact aponeurosis tethers the torn tendon stump (Figures 3A-3C). 

Chronic complete tears demonstrate heterogeneous signal intensity and fluid signal at the tendon, as well as muscle belly atrophy.⁹ A small percentage of distal biceps brachii tendons are bifid ¹⁰ (Figure 4). 

When injured, 75% have complete rupture of the short head with 17% of these having additional complete rupture of the long head, whereas 50% of those with complete rupture of the short head have partial tear or tendinosis of the long head.

Continue to: Partial distal bicep tears...

 

Partial distal bicep tears are characterized on MRI by focal or partial detachment of the tendon at the radial tuberosity with fluid filling the site of the tear. The degree of partial tearing can be assessed on MRI (Figures 5A, 5B). 

In distal biceps tendinosis, increased signals of thickened tendon fibers at the radial tuberosity are evident without focal discontinuity^7,8 (Figures 6A-6C). Patients may display attenuation of the distal tendon fibers or adjacent fluid distension representing bicipitoradial bursitis (Figures 7A, 7B).

MRI is useful in assessing the distal biceps tendon in the postoperative setting to evaluate the integrity of a repaired tendon. Cortical fixation button technique for repair creates minimal susceptibility artifacts on MRI. Postoperative MRI typically demonstrates a transverse hole drilled through the proximal radius at the site of the tuberosity with a cortical fixation button flush against the posterior radial cortex (Figures 8A-8D). 

The postoperative complication of heterotopic ossification can occasionally be observed on plain radiograph at the site of surgery, but it is less common with the current surgical technique than in the past.¹¹

ABSTRACT

Lesser tuberosity osteotomy (LTO) and subscapularis tenotomy (ST) are used for takedown of the subscapularis during shoulder arthroplasty. LTO offers the theoretical but unproven benefit of improved healing and function of the subscapularis. However, humeral stem subsidence and loosening may be greater when osteotomy is performed, which may compromise functional outcomes. Our hypothesis is that no difference in proximal collar press-fit humeral stem subsidence or loosening exists, with no impairment of functional outcomes using the LTO technique.

During the surgical approach for total shoulder arthroplasty (TSA), the subscapularis is taken down for adequate exposure to the glenohumeral joint. Various methods are available for taking down the subscapularis, including lesser tuberosity osteotomy (LTO) and a subscapularis tenotomy (ST). LTO offers the theoretical but unproven benefit of improved healing and function of the subscapularis secondary to bone-to-bone healing. One concern, however, is that humeral stem subsidence may be greater when an osteotomy is performed owing to compromise of metaphyseal cortical bone, which may compromise functional outcomes. The humeral stem design may also influence subsidence when metaphyseal bone proximally is compromised. This is a concern in both metaphyseal and diaphyseal fitting stems. Metaphyseal collars on diaphyseal fitting stems rely on adequate bone stock in the metaphysis to provide the additional support needed. Also, posterior subluxation remains a challenge in shoulder arthroplasty. The integrity of the subscapularis is important in prevention of posterior subluxation.¹ To our knowledge, no study to date has directly compared differences in humeral stem subsidence, loosening, or posterior subluxation between LTO and ST techniques with any humeral stem design. Our hypothesis is that no difference in proximal collar press-fit humeral stem subsidence or loosening exists, with no impairment of functional outcomes using the LTO technique. We also hypothesize that no difference in posterior subluxation exists between LTO and ST techniques.

MATERIALS AND METHODS

INCLUSION CRITERIA

Consecutive patients with a minimum of 12 months of radiographic follow-up were selected from 2007 to 2010 after TSA was performed by 1 of the senior authors (Dr. Miller and Dr. Voloshin). Study patients underwent primary TSA for primary osteoarthritis or rheumatoid arthritis.

EXCLUSION CRITERIA

Patients were excluded if they underwent TSA for posttraumatic glenohumeral arthritis, hemiarthroplasty, or osteonecrosis. Patients were also excluded if a rotator cuff tear was discovered intraoperatively or if they had a history of a rotator cuff repair. Additional exclusion criteria included postoperative trauma to the operative shoulder, postoperative infection, extensive documentation of chronic pain, and underlying neurologic disorder (eg, Parkinson disease, dystonia). Patients with a history of diabetes mellitus were not excluded.

SURGICAL TECHNIQUE

All patients underwent TSA via a deltopectoral approach in a modified beach chair position. Biceps tendons were tenodesed at the level of the pectoralis major. All patients received the same proximal collar press-fit implant (Bigliani-Flatow; Zimmer Biomet). These stems provide rotational stability in the metaphyseal segment via fins, vertical stability with the proximal collar, and distal fixation via an interference fit. All parts of the procedure were performed in similar fashion with the exception of ST vs LTO (Figures 1A-1D). 

All patients followed the same postoperative rehabilitation protocol.

Continue to: LTO was performed as the primary...

 

LESSER TUBEROSITY OSTEOTOMY

LTO was performed as the primary or preferred technique of 1 surgeon. After completion of the biceps tenodesis, the lesser tuberosity is reflected off with the subscapularis intact using an osteotome. After placement of the press-fit humeral stem, the LTO is repaired using No. 5 Ethibond Excel sutures (Ethicon) passed through previously created bone tunnels in the greater tuberosity. These sutures are tied over metal buttons over the lateral cortex of the greater tuberosity. Last, the lateral corner of the rotator interval is repaired using a single No. 2 FiberWire (Arthrex).²

SUBSCAPULARIS TENOTOMY

ST is the preferred surgical technique of the second surgeon. After a biceps tenodesis, the subscapularis tendon is released from the lesser tuberosity at the margin of the bicipital groove. Through careful dissection, a single flap including the underlying capsule is created and reflected medially to the level of the coracoid. After placement of the press-fit humeral stem and humeral head, the subscapularis is repaired back in place through previous bone tunnels and with a No. 5 Ethibond Excel suture under the appropriate tension. Then, the lateral corner of the rotator interval is closed using a single No. 2 Ethibond Excel suture in a figure-of-eight fashion.²

RADIOGRAPHIC ANALYSIS

The primary variables analyzed were subsidence and loosening. Additional variables, including humeral-acromial distance (HAD) and subluxation index, were also analyzed to assess for any additional impact caused by subsidence or loosening.³ All radiographic measurements were taken from the Grashey (true anteroposterior) view, except subluxation index, which was calculated using the axillary view. All radiographic measurements were completed by 3 independent reviewers. All radiographs were completed in a consistent manner according to postoperative protocols.

HAD was measured preoperatively, immediately postoperatively, and at final follow-up at a minimum of 1 year. The HAD was measured from the lowest point on the acromion to the humerus using a perpendicular line (Figure 2).

Subsidence of the prosthesis was calculated by determining the difference between immediate postoperative heights of the prosthesis in comparison to the value of the final follow-up films. To calculate the height, 2 lines were drawn, 1 line was drawn perpendicular to the top of the prosthetic head and 1 perpendicular to the top of the greater tuberosity (Figure 3). 

This distance was measured both immediately postoperatively and at final follow-up. The height difference between the 2 time points was used to calculate the subsidence.

Continue to: Posterior subluxation is indicated...

 

Posterior subluxation is indicated by a value >65%, a centered head is between 35% and 65%, and anterior subluxation is indicated by a value <35% (Figure 4).³ 

The subluxation index was calculated using the method outlined by Gerber and colleagues³ using the axillary view radiograph (Figure 4). The midpoint of the glenoid was identified. Then, a perpendicular line (line z) was drawn projecting from the midpoint of the glenoid through the humeral head. Two lines were then drawn in an anterior-to-posterior direction on the humeral head. The first line (measurement A) was drawn from the posterior cortex to the previous perpendicular line (line z). The second line (measurement B) was drawn from the posterior to the anterior cortex. Measurements A and B should be calculated at the same height on the humeral head, roughly at the anatomic neck. The subluxation index is then calculated as A/B × 100. This was calculated both preoperatively and postoperatively.

The humeral stems were evaluated for loosening by assessing for lucency on final radiographic follow-up films. These were evaluated in a zonal fashion as demonstrated by Sanchez-Sotelo and colleagues⁴ and in Figure 5. 

The humeral stem-bone interface was evaluated in 8 radiographic zones. A lucency was not considered significant unless >2 mm. The zone of lucency was then determined for any significant lucencies. Zones 1 and 7 were at the level of the neck component, zones 2 and 6 were at the proximal half of the stem component, and zones 3 and 5 were at the distal half of the stem component. Zone 8 was noted directly inferior to the humeral head prosthesis.⁴

FUNCTIONAL OUTCOME EVALUATION

Before clinical evaluation, each study patient completed the Western Ontario Osteoarthritis of the Shoulder (WOOS) index; the Disabilities of the Hand, Arm and Shoulder (DASH) questionnaire, and the pain and function sections of the Constant score. The functional outcomes scores were captured postoperatively from October to November 2011. The WOOS is a validated outcome measure specific to osteoarthritis of the shoulder and has been used in prior studies evaluating outcomes of TSA.^5-7 Previous studies have determined that the minimal clinically important difference for the WOOS score is 15 on a normalized 0 to 100 scale (100 being the best). The DASH score is a validated outcome measure for disorders of the upper extremity but is not specific to osteoarthritis of the shoulder.⁸ The Constant score is a validated outcome measure for a number of shoulder disorders, including TSA.^9,10

STATISTICAL ANALYSIS

Statistical analyses were completed by a trained biostatistician. A power analysis was calculated using the noninferiority test to determine if adequate data had been obtained for this study. This was calculated by using previously accepted data demonstrating a statistically significant difference for subsidence and HAD. The data from these studies were used to make assumptions regarding accepted standard deviations and noninferiority margins, as calculated from the mean values of the 2 groups analyzed in each respective study.^4,11 This analysis demonstrated power of 0.97 and 0.85 for the subsidence and HAD, respectively, given the current sample sizes. Intraclass coefficients were calculated to evaluate the measurements obtained during the radiographic analysis to determine the interrater agreement. Two samples’ t tests were calculated for the variables analyzed, along with P values and means.

RESULTS

DEMOGRAPHICS

A total of 51 consecutive patients were retrospectively selected for analysis. Of these, 16 patients were excluded from the study because they had <9 months of radiographic follow-up and were unavailable for further follow-up evaluation. Of the remaining 35 patients available for analysis, 4 patients had bilateral TSA, providing 39 shoulders for evaluation. Demographic characteristics of the study cohort are reported in Table 1.

Table 1. Demographic Characteristics
	Tenotomy (n = 24)	Osteotomy (n = 15)	P-value
Age	68.2 [7.4]	70.2 [7.1]	0.46
Follow-up	20.6 [11.5]	18.5 [6.25]	0.94
Females	7 (29%)	6 (40%)	0.58
Dominant shoulder	14 (58%)	8 (53%)	0.81
Primary Diagnosis
Osteoarthritis	22 (92%)	15 (100%)
Rheumatoid arthritis	2 (8%)	0 (0%)

  Fifteen patients underwent LTO, and 24 underwent ST. One patient underwent a tenotomy of the right shoulder and LTO of the left shoulder. Three LTOs were performed by the surgeon who primarily performed ST, owing to potential benefits of LTO. He eventually returned to his preferred technique of ST because of surgeon preference. Three ST procedures were completed by the surgeon who typically performed LTO at the start of the series prior to establishing LTO as his preferred technique. There was no significant difference between the study populations in terms of age, follow-up, male-to-female ratio, hand dominance, and primary diagnosis of osteoarthritis vs rheumatoid arthritis.

Continue to: There was no significant difference...

 

RADIOGRAPHIC DATA

There was no significant difference in preoperative HAD between the LTO and ST groups (9.5 ± 2.4 mm vs 10.9 ± 2.7 mm, P = .11). The immediate postoperative HAD was statistically significant between the LTO and ST groups (11.9 ± 3.7 mm vs 15.9 ± 4.5 mm, P = .005). There was as statistically significant difference noted in the final follow-up films between the LTO and ST groups (11.8 ± 3.2 mm vs 14.5 ± 3.9 mm, P = .025) (Table 2). 

Table 2. Radiographic Data
Humeral Acromial Distance
	LTO		ST		P-Value
Preoperative, mm	9.5	[2.4]	10.9	[2.7]	0.11
Postoperative, mm	11.9	[3.7]	15.9	[4.5]	0.005
Final follow-up, mm	11.8	[3.2]	14.5	[3.9]	0.025
Subsidence
	LTO		ST		P-Value
Subsidence, mm	2.8	[3.1]	2.5	[3.1]	0.72
Subluxation Index
	LTO		ST		P-Value
Preoperative, %	0.55	[0.06]	0.54	[0.07]	0.45
Postoperative, %	0.55	[0.09]	0.48	[0.05]	0.015
Lucent Lines
	LTO		ST		P-Value
Lines >2 mm, %	0.00		0.08		0.51

Abbreviations: LTO, lesser tuberosity osteotomy; ST, subscapularis tenotomy.

There were no statistically significant differences found in subsidence between LTO and ST groups at final follow-up (2.8 mm ± 3.1 mm vs 2.5 mm ± 3.1 mm, P = .72) (Table 2). No statistically significant difference was noted in the subluxation index between the LTO and ST groups (0.55% ± .06% vs 0.54% ± 0.07%, P = .45), but there was a statistically significant difference noted postoperatively between the LTO and ST groups (0.55% ± 0.09% vs .48% ± 0.05%, P = .015) (Table 2).

Two stems were noted to have lucent lines >2 mm, both within the ST cohort. Each had 1 stem zone >2 mm, 1 in zone 7, and 1 in zone 4. No statistically significant difference was identified between the LTO and ST groups (0/15 vs 2/24, P = .51) (Table 2).

FUNCTIONAL OUTCOMES

Study patients were evaluated using functional outcome scores, including the Constant, WOOS, and DASH scores (Table 3). 

Table 3. Functional Data
	LTO		ST		P-Value
WOOS index	93.3	[5.3]	81.5	[20.8]	0.013
DASH score	8.4	[6.6]	13.8	[4.9]	0.13
Constant score	83.3	[9.1]	81.8	[10.1]	0.64

Abbreviations: DASH, disabilities of the arm, shoulder and hand; WOOS, Western Ontario Osteoarthritis of the Shoulder.

No statistically significant differences were noted in the DASH scores (8.4 ± 6.6 vs 13.8 ± 4.9, P = .13) or Constant scores (83.3 ± 9.1 vs 81.8 ± 10.1, P = .64) between the LTO and ST cohorts. There was a statistically significant difference between the WOOS scores (93.3 ± 5.3 vs 81.5 ± 20.8, P = .013). Because separate radiographic reviews were done by 3 independent personnel at 3 different times, it was important to ensure agreement among the reviewers. This was compared using the intraclass correlation coefficients. In the statistical analysis completed, the intraclass coefficients showed the 3 reviewers agreed with each other throughout the radiographic analysis (Table 4).

Table 4. Testing Agreement: ICC
		ICC	CI, 2.5%	CI, 97.5%
HAD	Preoperative	0.4451	0.2202	0.6443
	Postoperative	0.6997	0.4836	0.834
	Final follow-up	0.5575	0.3592	0.7218
Subsidence		0.6863	0.5349	0.807
SI	Preoperative	0.3087	0.1061	0.5213
	Final follow-up	0.5364	0.299	0.7186

Abbreviations: CI, confidence interval; HAD, humeral acromial distance; ICC, intraclass correlation coefficient; SI, subluxation index.

DISCUSSION

At final follow-up, we identified no statistically significant difference between the LTO and ST patients in subsidence, lucent lines >2 mm, or functional outcomes (Constant and DASH scores) in patients who underwent TSA with the same proximal collar press-fit humeral stem. In regard to the functional outcome scores, although the WOOS score was statistically significant (P = .013) between the LTO and ST cohorts, we do not feel that this is clinically relevant because it does not reach the minimal clinically important difference threshold of 15 points.⁸

A statistically significant difference was noted in postoperative subluxation index but was not clinically relevant, because the values between the LTO and ST groups (0.55 vs 0.48) still showed a centered humeral head. Gerber and colleagues³ discussed using a value of 0.65 as a measure of posterior humeral head subluxation, whereas Walch and colleagues¹² defined posterior humeral head subluxation as a value >0.55. On the basis of these numbers, the values obtained in this study demonstrated that the postoperative values were still centered on the glenoid, and therefore were not clinically significant.^3,12

Continue to: In regard to HAD, there...

 

In regard to HAD, there was a statistically significant difference noted postoperatively (P = .005) and at final follow-up (P = .025) between the LTO and ST cohorts. Saupe and colleagues¹³ demonstrated that a HAD <7 mm was considered abnormal and reflected subacromial space narrowing. The values noted in the LTO and ST patients on postoperative and final follow-up radiographs were statistically significant (Table 2), but not clinically relevant because both were >7 mm. A potential source for the variation in HAD may be due to X-ray position and angle.

Studies have shown a concern regarding the integrity of the subscapularis after tenotomy or peel used in TSA with abnormal subscapularis function.^14,15 Miller and colleagues¹⁵ reported 41 patients, nearly two-thirds, of whom described subscapularis dysfunction. Those authors’ response to the poor clinical outcomes was to remove a fleck of bone with the tendon to achieve “bone-to-bone” healing.¹⁴ Gerber and colleagues¹⁶ reported on a series of patients using LTO and repair in TSA with 75% and 89% intact subscapularis function on clinical testing.¹⁶ Studies by Qureshi and colleagues¹⁷ and Scalise and colleagues¹⁸ showed similar results after LTO. Biomechanical studies have shown mixed results. Ponce and colleagues¹⁹ showed biomechanically superior results for LTO in comparison to the various repair techniques for ST. In another study, Giuseffi and colleagues²⁰ showed no difference in LTO vs ST during biomechanical testing. In response to the increased concern regarding subscapularis integrity, Caplan and colleagues²¹ reported on 45 arthroplasties in 43 patients with improved postoperative testing with intact subscapularis testing in 90% to 100% of patients. A level 1 randomized control trial conducted by Lapner and colleagues²² did not demonstrate any clear clinical advantage of LTO vs ST. Controversy still exists regarding which is the preferred technique for TSA.

Sanchez-Sotelo and colleagues⁴ evaluated uncemented humeral components in 72 patients who underwent TSA. They found a humeral component was at risk for loosening if a radiolucent line ≥2 mm was present in at least 3 radiographic zones. They also evaluated tilt or subsidence by measurement and whether the components were observed to have changed. Their measured values correlated with their observed values. That study provided a benchmark for evaluation of loosening and subsidence used during this study.⁴ Although radiographic follow-up is limited in this study, we feel that any potential subsidence secondary to use of the LTO technique would be radiographically apparent at 1 year. There were 16 patients without adequate radiographic follow-up included in the study. However, we feel that this was not a large concern, because the study was adequately powered with the patients available to determine a difference based on subsidence.

CONCLUSION

We found no difference in subsidence, lucent lines >2 mm, posterior subluxation, and the Constant and DASH functional outcome scores when we compared TSA performed by a LTO with an ST technique with proximal collar press-fit humeral stem. These data cannot be extrapolated to metaphyseal fit stems, which may exhibit different settling characteristics in the setting of the LTO technique.

This paper will be judged for the Resident Writer’s Award.

ABSTRACT

Lesser tuberosity osteotomy (LTO) and subscapularis tenotomy (ST) are used for takedown of the subscapularis during shoulder arthroplasty. LTO offers the theoretical but unproven benefit of improved healing and function of the subscapularis. However, humeral stem subsidence and loosening may be greater when osteotomy is performed, which may compromise functional outcomes. Our hypothesis is that no difference in proximal collar press-fit humeral stem subsidence or loosening exists, with no impairment of functional outcomes using the LTO technique.

During the surgical approach for total shoulder arthroplasty (TSA), the subscapularis is taken down for adequate exposure to the glenohumeral joint. Various methods are available for taking down the subscapularis, including lesser tuberosity osteotomy (LTO) and a subscapularis tenotomy (ST). LTO offers the theoretical but unproven benefit of improved healing and function of the subscapularis secondary to bone-to-bone healing. One concern, however, is that humeral stem subsidence may be greater when an osteotomy is performed owing to compromise of metaphyseal cortical bone, which may compromise functional outcomes. The humeral stem design may also influence subsidence when metaphyseal bone proximally is compromised. This is a concern in both metaphyseal and diaphyseal fitting stems. Metaphyseal collars on diaphyseal fitting stems rely on adequate bone stock in the metaphysis to provide the additional support needed. Also, posterior subluxation remains a challenge in shoulder arthroplasty. The integrity of the subscapularis is important in prevention of posterior subluxation.¹ To our knowledge, no study to date has directly compared differences in humeral stem subsidence, loosening, or posterior subluxation between LTO and ST techniques with any humeral stem design. Our hypothesis is that no difference in proximal collar press-fit humeral stem subsidence or loosening exists, with no impairment of functional outcomes using the LTO technique. We also hypothesize that no difference in posterior subluxation exists between LTO and ST techniques.

MATERIALS AND METHODS

INCLUSION CRITERIA

Consecutive patients with a minimum of 12 months of radiographic follow-up were selected from 2007 to 2010 after TSA was performed by 1 of the senior authors (Dr. Miller and Dr. Voloshin). Study patients underwent primary TSA for primary osteoarthritis or rheumatoid arthritis.

EXCLUSION CRITERIA

Patients were excluded if they underwent TSA for posttraumatic glenohumeral arthritis, hemiarthroplasty, or osteonecrosis. Patients were also excluded if a rotator cuff tear was discovered intraoperatively or if they had a history of a rotator cuff repair. Additional exclusion criteria included postoperative trauma to the operative shoulder, postoperative infection, extensive documentation of chronic pain, and underlying neurologic disorder (eg, Parkinson disease, dystonia). Patients with a history of diabetes mellitus were not excluded.

SURGICAL TECHNIQUE

All patients underwent TSA via a deltopectoral approach in a modified beach chair position. Biceps tendons were tenodesed at the level of the pectoralis major. All patients received the same proximal collar press-fit implant (Bigliani-Flatow; Zimmer Biomet). These stems provide rotational stability in the metaphyseal segment via fins, vertical stability with the proximal collar, and distal fixation via an interference fit. All parts of the procedure were performed in similar fashion with the exception of ST vs LTO (Figures 1A-1D). 

All patients followed the same postoperative rehabilitation protocol.

Continue to: LTO was performed as the primary...

 

LESSER TUBEROSITY OSTEOTOMY

LTO was performed as the primary or preferred technique of 1 surgeon. After completion of the biceps tenodesis, the lesser tuberosity is reflected off with the subscapularis intact using an osteotome. After placement of the press-fit humeral stem, the LTO is repaired using No. 5 Ethibond Excel sutures (Ethicon) passed through previously created bone tunnels in the greater tuberosity. These sutures are tied over metal buttons over the lateral cortex of the greater tuberosity. Last, the lateral corner of the rotator interval is repaired using a single No. 2 FiberWire (Arthrex).²

SUBSCAPULARIS TENOTOMY

ST is the preferred surgical technique of the second surgeon. After a biceps tenodesis, the subscapularis tendon is released from the lesser tuberosity at the margin of the bicipital groove. Through careful dissection, a single flap including the underlying capsule is created and reflected medially to the level of the coracoid. After placement of the press-fit humeral stem and humeral head, the subscapularis is repaired back in place through previous bone tunnels and with a No. 5 Ethibond Excel suture under the appropriate tension. Then, the lateral corner of the rotator interval is closed using a single No. 2 Ethibond Excel suture in a figure-of-eight fashion.²

RADIOGRAPHIC ANALYSIS

The primary variables analyzed were subsidence and loosening. Additional variables, including humeral-acromial distance (HAD) and subluxation index, were also analyzed to assess for any additional impact caused by subsidence or loosening.³ All radiographic measurements were taken from the Grashey (true anteroposterior) view, except subluxation index, which was calculated using the axillary view. All radiographic measurements were completed by 3 independent reviewers. All radiographs were completed in a consistent manner according to postoperative protocols.

HAD was measured preoperatively, immediately postoperatively, and at final follow-up at a minimum of 1 year. The HAD was measured from the lowest point on the acromion to the humerus using a perpendicular line (Figure 2).

Subsidence of the prosthesis was calculated by determining the difference between immediate postoperative heights of the prosthesis in comparison to the value of the final follow-up films. To calculate the height, 2 lines were drawn, 1 line was drawn perpendicular to the top of the prosthetic head and 1 perpendicular to the top of the greater tuberosity (Figure 3). 

This distance was measured both immediately postoperatively and at final follow-up. The height difference between the 2 time points was used to calculate the subsidence.

Continue to: Posterior subluxation is indicated...

 

Posterior subluxation is indicated by a value >65%, a centered head is between 35% and 65%, and anterior subluxation is indicated by a value <35% (Figure 4).³ 

The subluxation index was calculated using the method outlined by Gerber and colleagues³ using the axillary view radiograph (Figure 4). The midpoint of the glenoid was identified. Then, a perpendicular line (line z) was drawn projecting from the midpoint of the glenoid through the humeral head. Two lines were then drawn in an anterior-to-posterior direction on the humeral head. The first line (measurement A) was drawn from the posterior cortex to the previous perpendicular line (line z). The second line (measurement B) was drawn from the posterior to the anterior cortex. Measurements A and B should be calculated at the same height on the humeral head, roughly at the anatomic neck. The subluxation index is then calculated as A/B × 100. This was calculated both preoperatively and postoperatively.

The humeral stems were evaluated for loosening by assessing for lucency on final radiographic follow-up films. These were evaluated in a zonal fashion as demonstrated by Sanchez-Sotelo and colleagues⁴ and in Figure 5. 

The humeral stem-bone interface was evaluated in 8 radiographic zones. A lucency was not considered significant unless >2 mm. The zone of lucency was then determined for any significant lucencies. Zones 1 and 7 were at the level of the neck component, zones 2 and 6 were at the proximal half of the stem component, and zones 3 and 5 were at the distal half of the stem component. Zone 8 was noted directly inferior to the humeral head prosthesis.⁴

FUNCTIONAL OUTCOME EVALUATION

Before clinical evaluation, each study patient completed the Western Ontario Osteoarthritis of the Shoulder (WOOS) index; the Disabilities of the Hand, Arm and Shoulder (DASH) questionnaire, and the pain and function sections of the Constant score. The functional outcomes scores were captured postoperatively from October to November 2011. The WOOS is a validated outcome measure specific to osteoarthritis of the shoulder and has been used in prior studies evaluating outcomes of TSA.^5-7 Previous studies have determined that the minimal clinically important difference for the WOOS score is 15 on a normalized 0 to 100 scale (100 being the best). The DASH score is a validated outcome measure for disorders of the upper extremity but is not specific to osteoarthritis of the shoulder.⁸ The Constant score is a validated outcome measure for a number of shoulder disorders, including TSA.^9,10

STATISTICAL ANALYSIS

Statistical analyses were completed by a trained biostatistician. A power analysis was calculated using the noninferiority test to determine if adequate data had been obtained for this study. This was calculated by using previously accepted data demonstrating a statistically significant difference for subsidence and HAD. The data from these studies were used to make assumptions regarding accepted standard deviations and noninferiority margins, as calculated from the mean values of the 2 groups analyzed in each respective study.^4,11 This analysis demonstrated power of 0.97 and 0.85 for the subsidence and HAD, respectively, given the current sample sizes. Intraclass coefficients were calculated to evaluate the measurements obtained during the radiographic analysis to determine the interrater agreement. Two samples’ t tests were calculated for the variables analyzed, along with P values and means.

RESULTS

DEMOGRAPHICS

A total of 51 consecutive patients were retrospectively selected for analysis. Of these, 16 patients were excluded from the study because they had <9 months of radiographic follow-up and were unavailable for further follow-up evaluation. Of the remaining 35 patients available for analysis, 4 patients had bilateral TSA, providing 39 shoulders for evaluation. Demographic characteristics of the study cohort are reported in Table 1.

Table 1. Demographic Characteristics
	Tenotomy (n = 24)	Osteotomy (n = 15)	P-value
Age	68.2 [7.4]	70.2 [7.1]	0.46
Follow-up	20.6 [11.5]	18.5 [6.25]	0.94
Females	7 (29%)	6 (40%)	0.58
Dominant shoulder	14 (58%)	8 (53%)	0.81
Primary Diagnosis
Osteoarthritis	22 (92%)	15 (100%)
Rheumatoid arthritis	2 (8%)	0 (0%)

  Fifteen patients underwent LTO, and 24 underwent ST. One patient underwent a tenotomy of the right shoulder and LTO of the left shoulder. Three LTOs were performed by the surgeon who primarily performed ST, owing to potential benefits of LTO. He eventually returned to his preferred technique of ST because of surgeon preference. Three ST procedures were completed by the surgeon who typically performed LTO at the start of the series prior to establishing LTO as his preferred technique. There was no significant difference between the study populations in terms of age, follow-up, male-to-female ratio, hand dominance, and primary diagnosis of osteoarthritis vs rheumatoid arthritis.

Continue to: There was no significant difference...

 

RADIOGRAPHIC DATA

There was no significant difference in preoperative HAD between the LTO and ST groups (9.5 ± 2.4 mm vs 10.9 ± 2.7 mm, P = .11). The immediate postoperative HAD was statistically significant between the LTO and ST groups (11.9 ± 3.7 mm vs 15.9 ± 4.5 mm, P = .005). There was as statistically significant difference noted in the final follow-up films between the LTO and ST groups (11.8 ± 3.2 mm vs 14.5 ± 3.9 mm, P = .025) (Table 2). 

Table 2. Radiographic Data
Humeral Acromial Distance
	LTO		ST		P-Value
Preoperative, mm	9.5	[2.4]	10.9	[2.7]	0.11
Postoperative, mm	11.9	[3.7]	15.9	[4.5]	0.005
Final follow-up, mm	11.8	[3.2]	14.5	[3.9]	0.025
Subsidence
	LTO		ST		P-Value
Subsidence, mm	2.8	[3.1]	2.5	[3.1]	0.72
Subluxation Index
	LTO		ST		P-Value
Preoperative, %	0.55	[0.06]	0.54	[0.07]	0.45
Postoperative, %	0.55	[0.09]	0.48	[0.05]	0.015
Lucent Lines
	LTO		ST		P-Value
Lines >2 mm, %	0.00		0.08		0.51

Abbreviations: LTO, lesser tuberosity osteotomy; ST, subscapularis tenotomy.

There were no statistically significant differences found in subsidence between LTO and ST groups at final follow-up (2.8 mm ± 3.1 mm vs 2.5 mm ± 3.1 mm, P = .72) (Table 2). No statistically significant difference was noted in the subluxation index between the LTO and ST groups (0.55% ± .06% vs 0.54% ± 0.07%, P = .45), but there was a statistically significant difference noted postoperatively between the LTO and ST groups (0.55% ± 0.09% vs .48% ± 0.05%, P = .015) (Table 2).

Two stems were noted to have lucent lines >2 mm, both within the ST cohort. Each had 1 stem zone >2 mm, 1 in zone 7, and 1 in zone 4. No statistically significant difference was identified between the LTO and ST groups (0/15 vs 2/24, P = .51) (Table 2).

FUNCTIONAL OUTCOMES

Study patients were evaluated using functional outcome scores, including the Constant, WOOS, and DASH scores (Table 3). 

Table 3. Functional Data
	LTO		ST		P-Value
WOOS index	93.3	[5.3]	81.5	[20.8]	0.013
DASH score	8.4	[6.6]	13.8	[4.9]	0.13
Constant score	83.3	[9.1]	81.8	[10.1]	0.64

Abbreviations: DASH, disabilities of the arm, shoulder and hand; WOOS, Western Ontario Osteoarthritis of the Shoulder.

No statistically significant differences were noted in the DASH scores (8.4 ± 6.6 vs 13.8 ± 4.9, P = .13) or Constant scores (83.3 ± 9.1 vs 81.8 ± 10.1, P = .64) between the LTO and ST cohorts. There was a statistically significant difference between the WOOS scores (93.3 ± 5.3 vs 81.5 ± 20.8, P = .013). Because separate radiographic reviews were done by 3 independent personnel at 3 different times, it was important to ensure agreement among the reviewers. This was compared using the intraclass correlation coefficients. In the statistical analysis completed, the intraclass coefficients showed the 3 reviewers agreed with each other throughout the radiographic analysis (Table 4).

Table 4. Testing Agreement: ICC
		ICC	CI, 2.5%	CI, 97.5%
HAD	Preoperative	0.4451	0.2202	0.6443
	Postoperative	0.6997	0.4836	0.834
	Final follow-up	0.5575	0.3592	0.7218
Subsidence		0.6863	0.5349	0.807
SI	Preoperative	0.3087	0.1061	0.5213
	Final follow-up	0.5364	0.299	0.7186

Abbreviations: CI, confidence interval; HAD, humeral acromial distance; ICC, intraclass correlation coefficient; SI, subluxation index.

DISCUSSION

At final follow-up, we identified no statistically significant difference between the LTO and ST patients in subsidence, lucent lines >2 mm, or functional outcomes (Constant and DASH scores) in patients who underwent TSA with the same proximal collar press-fit humeral stem. In regard to the functional outcome scores, although the WOOS score was statistically significant (P = .013) between the LTO and ST cohorts, we do not feel that this is clinically relevant because it does not reach the minimal clinically important difference threshold of 15 points.⁸

A statistically significant difference was noted in postoperative subluxation index but was not clinically relevant, because the values between the LTO and ST groups (0.55 vs 0.48) still showed a centered humeral head. Gerber and colleagues³ discussed using a value of 0.65 as a measure of posterior humeral head subluxation, whereas Walch and colleagues¹² defined posterior humeral head subluxation as a value >0.55. On the basis of these numbers, the values obtained in this study demonstrated that the postoperative values were still centered on the glenoid, and therefore were not clinically significant.^3,12

Continue to: In regard to HAD, there...

 

In regard to HAD, there was a statistically significant difference noted postoperatively (P = .005) and at final follow-up (P = .025) between the LTO and ST cohorts. Saupe and colleagues¹³ demonstrated that a HAD <7 mm was considered abnormal and reflected subacromial space narrowing. The values noted in the LTO and ST patients on postoperative and final follow-up radiographs were statistically significant (Table 2), but not clinically relevant because both were >7 mm. A potential source for the variation in HAD may be due to X-ray position and angle.

Studies have shown a concern regarding the integrity of the subscapularis after tenotomy or peel used in TSA with abnormal subscapularis function.^14,15 Miller and colleagues¹⁵ reported 41 patients, nearly two-thirds, of whom described subscapularis dysfunction. Those authors’ response to the poor clinical outcomes was to remove a fleck of bone with the tendon to achieve “bone-to-bone” healing.¹⁴ Gerber and colleagues¹⁶ reported on a series of patients using LTO and repair in TSA with 75% and 89% intact subscapularis function on clinical testing.¹⁶ Studies by Qureshi and colleagues¹⁷ and Scalise and colleagues¹⁸ showed similar results after LTO. Biomechanical studies have shown mixed results. Ponce and colleagues¹⁹ showed biomechanically superior results for LTO in comparison to the various repair techniques for ST. In another study, Giuseffi and colleagues²⁰ showed no difference in LTO vs ST during biomechanical testing. In response to the increased concern regarding subscapularis integrity, Caplan and colleagues²¹ reported on 45 arthroplasties in 43 patients with improved postoperative testing with intact subscapularis testing in 90% to 100% of patients. A level 1 randomized control trial conducted by Lapner and colleagues²² did not demonstrate any clear clinical advantage of LTO vs ST. Controversy still exists regarding which is the preferred technique for TSA.

Sanchez-Sotelo and colleagues⁴ evaluated uncemented humeral components in 72 patients who underwent TSA. They found a humeral component was at risk for loosening if a radiolucent line ≥2 mm was present in at least 3 radiographic zones. They also evaluated tilt or subsidence by measurement and whether the components were observed to have changed. Their measured values correlated with their observed values. That study provided a benchmark for evaluation of loosening and subsidence used during this study.⁴ Although radiographic follow-up is limited in this study, we feel that any potential subsidence secondary to use of the LTO technique would be radiographically apparent at 1 year. There were 16 patients without adequate radiographic follow-up included in the study. However, we feel that this was not a large concern, because the study was adequately powered with the patients available to determine a difference based on subsidence.

CONCLUSION

We found no difference in subsidence, lucent lines >2 mm, posterior subluxation, and the Constant and DASH functional outcome scores when we compared TSA performed by a LTO with an ST technique with proximal collar press-fit humeral stem. These data cannot be extrapolated to metaphyseal fit stems, which may exhibit different settling characteristics in the setting of the LTO technique.

This paper will be judged for the Resident Writer’s Award.

ABSTRACT

Lesser tuberosity osteotomy (LTO) and subscapularis tenotomy (ST) are used for takedown of the subscapularis during shoulder arthroplasty. LTO offers the theoretical but unproven benefit of improved healing and function of the subscapularis. However, humeral stem subsidence and loosening may be greater when osteotomy is performed, which may compromise functional outcomes. Our hypothesis is that no difference in proximal collar press-fit humeral stem subsidence or loosening exists, with no impairment of functional outcomes using the LTO technique.

During the surgical approach for total shoulder arthroplasty (TSA), the subscapularis is taken down for adequate exposure to the glenohumeral joint. Various methods are available for taking down the subscapularis, including lesser tuberosity osteotomy (LTO) and a subscapularis tenotomy (ST). LTO offers the theoretical but unproven benefit of improved healing and function of the subscapularis secondary to bone-to-bone healing. One concern, however, is that humeral stem subsidence may be greater when an osteotomy is performed owing to compromise of metaphyseal cortical bone, which may compromise functional outcomes. The humeral stem design may also influence subsidence when metaphyseal bone proximally is compromised. This is a concern in both metaphyseal and diaphyseal fitting stems. Metaphyseal collars on diaphyseal fitting stems rely on adequate bone stock in the metaphysis to provide the additional support needed. Also, posterior subluxation remains a challenge in shoulder arthroplasty. The integrity of the subscapularis is important in prevention of posterior subluxation.¹ To our knowledge, no study to date has directly compared differences in humeral stem subsidence, loosening, or posterior subluxation between LTO and ST techniques with any humeral stem design. Our hypothesis is that no difference in proximal collar press-fit humeral stem subsidence or loosening exists, with no impairment of functional outcomes using the LTO technique. We also hypothesize that no difference in posterior subluxation exists between LTO and ST techniques.

MATERIALS AND METHODS

INCLUSION CRITERIA

Consecutive patients with a minimum of 12 months of radiographic follow-up were selected from 2007 to 2010 after TSA was performed by 1 of the senior authors (Dr. Miller and Dr. Voloshin). Study patients underwent primary TSA for primary osteoarthritis or rheumatoid arthritis.

EXCLUSION CRITERIA

Patients were excluded if they underwent TSA for posttraumatic glenohumeral arthritis, hemiarthroplasty, or osteonecrosis. Patients were also excluded if a rotator cuff tear was discovered intraoperatively or if they had a history of a rotator cuff repair. Additional exclusion criteria included postoperative trauma to the operative shoulder, postoperative infection, extensive documentation of chronic pain, and underlying neurologic disorder (eg, Parkinson disease, dystonia). Patients with a history of diabetes mellitus were not excluded.

SURGICAL TECHNIQUE

All patients underwent TSA via a deltopectoral approach in a modified beach chair position. Biceps tendons were tenodesed at the level of the pectoralis major. All patients received the same proximal collar press-fit implant (Bigliani-Flatow; Zimmer Biomet). These stems provide rotational stability in the metaphyseal segment via fins, vertical stability with the proximal collar, and distal fixation via an interference fit. All parts of the procedure were performed in similar fashion with the exception of ST vs LTO (Figures 1A-1D). 

All patients followed the same postoperative rehabilitation protocol.

Continue to: LTO was performed as the primary...

 

LESSER TUBEROSITY OSTEOTOMY

LTO was performed as the primary or preferred technique of 1 surgeon. After completion of the biceps tenodesis, the lesser tuberosity is reflected off with the subscapularis intact using an osteotome. After placement of the press-fit humeral stem, the LTO is repaired using No. 5 Ethibond Excel sutures (Ethicon) passed through previously created bone tunnels in the greater tuberosity. These sutures are tied over metal buttons over the lateral cortex of the greater tuberosity. Last, the lateral corner of the rotator interval is repaired using a single No. 2 FiberWire (Arthrex).²

SUBSCAPULARIS TENOTOMY

ST is the preferred surgical technique of the second surgeon. After a biceps tenodesis, the subscapularis tendon is released from the lesser tuberosity at the margin of the bicipital groove. Through careful dissection, a single flap including the underlying capsule is created and reflected medially to the level of the coracoid. After placement of the press-fit humeral stem and humeral head, the subscapularis is repaired back in place through previous bone tunnels and with a No. 5 Ethibond Excel suture under the appropriate tension. Then, the lateral corner of the rotator interval is closed using a single No. 2 Ethibond Excel suture in a figure-of-eight fashion.²

RADIOGRAPHIC ANALYSIS

The primary variables analyzed were subsidence and loosening. Additional variables, including humeral-acromial distance (HAD) and subluxation index, were also analyzed to assess for any additional impact caused by subsidence or loosening.³ All radiographic measurements were taken from the Grashey (true anteroposterior) view, except subluxation index, which was calculated using the axillary view. All radiographic measurements were completed by 3 independent reviewers. All radiographs were completed in a consistent manner according to postoperative protocols.

HAD was measured preoperatively, immediately postoperatively, and at final follow-up at a minimum of 1 year. The HAD was measured from the lowest point on the acromion to the humerus using a perpendicular line (Figure 2).

Subsidence of the prosthesis was calculated by determining the difference between immediate postoperative heights of the prosthesis in comparison to the value of the final follow-up films. To calculate the height, 2 lines were drawn, 1 line was drawn perpendicular to the top of the prosthetic head and 1 perpendicular to the top of the greater tuberosity (Figure 3). 

This distance was measured both immediately postoperatively and at final follow-up. The height difference between the 2 time points was used to calculate the subsidence.

Continue to: Posterior subluxation is indicated...

 

Posterior subluxation is indicated by a value >65%, a centered head is between 35% and 65%, and anterior subluxation is indicated by a value <35% (Figure 4).³ 

The subluxation index was calculated using the method outlined by Gerber and colleagues³ using the axillary view radiograph (Figure 4). The midpoint of the glenoid was identified. Then, a perpendicular line (line z) was drawn projecting from the midpoint of the glenoid through the humeral head. Two lines were then drawn in an anterior-to-posterior direction on the humeral head. The first line (measurement A) was drawn from the posterior cortex to the previous perpendicular line (line z). The second line (measurement B) was drawn from the posterior to the anterior cortex. Measurements A and B should be calculated at the same height on the humeral head, roughly at the anatomic neck. The subluxation index is then calculated as A/B × 100. This was calculated both preoperatively and postoperatively.

The humeral stems were evaluated for loosening by assessing for lucency on final radiographic follow-up films. These were evaluated in a zonal fashion as demonstrated by Sanchez-Sotelo and colleagues⁴ and in Figure 5. 

The humeral stem-bone interface was evaluated in 8 radiographic zones. A lucency was not considered significant unless >2 mm. The zone of lucency was then determined for any significant lucencies. Zones 1 and 7 were at the level of the neck component, zones 2 and 6 were at the proximal half of the stem component, and zones 3 and 5 were at the distal half of the stem component. Zone 8 was noted directly inferior to the humeral head prosthesis.⁴

FUNCTIONAL OUTCOME EVALUATION

Before clinical evaluation, each study patient completed the Western Ontario Osteoarthritis of the Shoulder (WOOS) index; the Disabilities of the Hand, Arm and Shoulder (DASH) questionnaire, and the pain and function sections of the Constant score. The functional outcomes scores were captured postoperatively from October to November 2011. The WOOS is a validated outcome measure specific to osteoarthritis of the shoulder and has been used in prior studies evaluating outcomes of TSA.^5-7 Previous studies have determined that the minimal clinically important difference for the WOOS score is 15 on a normalized 0 to 100 scale (100 being the best). The DASH score is a validated outcome measure for disorders of the upper extremity but is not specific to osteoarthritis of the shoulder.⁸ The Constant score is a validated outcome measure for a number of shoulder disorders, including TSA.^9,10

STATISTICAL ANALYSIS

Statistical analyses were completed by a trained biostatistician. A power analysis was calculated using the noninferiority test to determine if adequate data had been obtained for this study. This was calculated by using previously accepted data demonstrating a statistically significant difference for subsidence and HAD. The data from these studies were used to make assumptions regarding accepted standard deviations and noninferiority margins, as calculated from the mean values of the 2 groups analyzed in each respective study.^4,11 This analysis demonstrated power of 0.97 and 0.85 for the subsidence and HAD, respectively, given the current sample sizes. Intraclass coefficients were calculated to evaluate the measurements obtained during the radiographic analysis to determine the interrater agreement. Two samples’ t tests were calculated for the variables analyzed, along with P values and means.

RESULTS

DEMOGRAPHICS

A total of 51 consecutive patients were retrospectively selected for analysis. Of these, 16 patients were excluded from the study because they had <9 months of radiographic follow-up and were unavailable for further follow-up evaluation. Of the remaining 35 patients available for analysis, 4 patients had bilateral TSA, providing 39 shoulders for evaluation. Demographic characteristics of the study cohort are reported in Table 1.

Table 1. Demographic Characteristics
	Tenotomy (n = 24)	Osteotomy (n = 15)	P-value
Age	68.2 [7.4]	70.2 [7.1]	0.46
Follow-up	20.6 [11.5]	18.5 [6.25]	0.94
Females	7 (29%)	6 (40%)	0.58
Dominant shoulder	14 (58%)	8 (53%)	0.81
Primary Diagnosis
Osteoarthritis	22 (92%)	15 (100%)
Rheumatoid arthritis	2 (8%)	0 (0%)

  Fifteen patients underwent LTO, and 24 underwent ST. One patient underwent a tenotomy of the right shoulder and LTO of the left shoulder. Three LTOs were performed by the surgeon who primarily performed ST, owing to potential benefits of LTO. He eventually returned to his preferred technique of ST because of surgeon preference. Three ST procedures were completed by the surgeon who typically performed LTO at the start of the series prior to establishing LTO as his preferred technique. There was no significant difference between the study populations in terms of age, follow-up, male-to-female ratio, hand dominance, and primary diagnosis of osteoarthritis vs rheumatoid arthritis.

Continue to: There was no significant difference...

 

RADIOGRAPHIC DATA

There was no significant difference in preoperative HAD between the LTO and ST groups (9.5 ± 2.4 mm vs 10.9 ± 2.7 mm, P = .11). The immediate postoperative HAD was statistically significant between the LTO and ST groups (11.9 ± 3.7 mm vs 15.9 ± 4.5 mm, P = .005). There was as statistically significant difference noted in the final follow-up films between the LTO and ST groups (11.8 ± 3.2 mm vs 14.5 ± 3.9 mm, P = .025) (Table 2). 

Table 2. Radiographic Data
Humeral Acromial Distance
	LTO		ST		P-Value
Preoperative, mm	9.5	[2.4]	10.9	[2.7]	0.11
Postoperative, mm	11.9	[3.7]	15.9	[4.5]	0.005
Final follow-up, mm	11.8	[3.2]	14.5	[3.9]	0.025
Subsidence
	LTO		ST		P-Value
Subsidence, mm	2.8	[3.1]	2.5	[3.1]	0.72
Subluxation Index
	LTO		ST		P-Value
Preoperative, %	0.55	[0.06]	0.54	[0.07]	0.45
Postoperative, %	0.55	[0.09]	0.48	[0.05]	0.015
Lucent Lines
	LTO		ST		P-Value
Lines >2 mm, %	0.00		0.08		0.51

Abbreviations: LTO, lesser tuberosity osteotomy; ST, subscapularis tenotomy.

There were no statistically significant differences found in subsidence between LTO and ST groups at final follow-up (2.8 mm ± 3.1 mm vs 2.5 mm ± 3.1 mm, P = .72) (Table 2). No statistically significant difference was noted in the subluxation index between the LTO and ST groups (0.55% ± .06% vs 0.54% ± 0.07%, P = .45), but there was a statistically significant difference noted postoperatively between the LTO and ST groups (0.55% ± 0.09% vs .48% ± 0.05%, P = .015) (Table 2).

Two stems were noted to have lucent lines >2 mm, both within the ST cohort. Each had 1 stem zone >2 mm, 1 in zone 7, and 1 in zone 4. No statistically significant difference was identified between the LTO and ST groups (0/15 vs 2/24, P = .51) (Table 2).

FUNCTIONAL OUTCOMES

Study patients were evaluated using functional outcome scores, including the Constant, WOOS, and DASH scores (Table 3). 

Table 3. Functional Data
	LTO		ST		P-Value
WOOS index	93.3	[5.3]	81.5	[20.8]	0.013
DASH score	8.4	[6.6]	13.8	[4.9]	0.13
Constant score	83.3	[9.1]	81.8	[10.1]	0.64

Abbreviations: DASH, disabilities of the arm, shoulder and hand; WOOS, Western Ontario Osteoarthritis of the Shoulder.

No statistically significant differences were noted in the DASH scores (8.4 ± 6.6 vs 13.8 ± 4.9, P = .13) or Constant scores (83.3 ± 9.1 vs 81.8 ± 10.1, P = .64) between the LTO and ST cohorts. There was a statistically significant difference between the WOOS scores (93.3 ± 5.3 vs 81.5 ± 20.8, P = .013). Because separate radiographic reviews were done by 3 independent personnel at 3 different times, it was important to ensure agreement among the reviewers. This was compared using the intraclass correlation coefficients. In the statistical analysis completed, the intraclass coefficients showed the 3 reviewers agreed with each other throughout the radiographic analysis (Table 4).

Table 4. Testing Agreement: ICC
		ICC	CI, 2.5%	CI, 97.5%
HAD	Preoperative	0.4451	0.2202	0.6443
	Postoperative	0.6997	0.4836	0.834
	Final follow-up	0.5575	0.3592	0.7218
Subsidence		0.6863	0.5349	0.807
SI	Preoperative	0.3087	0.1061	0.5213
	Final follow-up	0.5364	0.299	0.7186

Abbreviations: CI, confidence interval; HAD, humeral acromial distance; ICC, intraclass correlation coefficient; SI, subluxation index.

DISCUSSION

At final follow-up, we identified no statistically significant difference between the LTO and ST patients in subsidence, lucent lines >2 mm, or functional outcomes (Constant and DASH scores) in patients who underwent TSA with the same proximal collar press-fit humeral stem. In regard to the functional outcome scores, although the WOOS score was statistically significant (P = .013) between the LTO and ST cohorts, we do not feel that this is clinically relevant because it does not reach the minimal clinically important difference threshold of 15 points.⁸

A statistically significant difference was noted in postoperative subluxation index but was not clinically relevant, because the values between the LTO and ST groups (0.55 vs 0.48) still showed a centered humeral head. Gerber and colleagues³ discussed using a value of 0.65 as a measure of posterior humeral head subluxation, whereas Walch and colleagues¹² defined posterior humeral head subluxation as a value >0.55. On the basis of these numbers, the values obtained in this study demonstrated that the postoperative values were still centered on the glenoid, and therefore were not clinically significant.^3,12

Continue to: In regard to HAD, there...

 

In regard to HAD, there was a statistically significant difference noted postoperatively (P = .005) and at final follow-up (P = .025) between the LTO and ST cohorts. Saupe and colleagues¹³ demonstrated that a HAD <7 mm was considered abnormal and reflected subacromial space narrowing. The values noted in the LTO and ST patients on postoperative and final follow-up radiographs were statistically significant (Table 2), but not clinically relevant because both were >7 mm. A potential source for the variation in HAD may be due to X-ray position and angle.

Studies have shown a concern regarding the integrity of the subscapularis after tenotomy or peel used in TSA with abnormal subscapularis function.^14,15 Miller and colleagues¹⁵ reported 41 patients, nearly two-thirds, of whom described subscapularis dysfunction. Those authors’ response to the poor clinical outcomes was to remove a fleck of bone with the tendon to achieve “bone-to-bone” healing.¹⁴ Gerber and colleagues¹⁶ reported on a series of patients using LTO and repair in TSA with 75% and 89% intact subscapularis function on clinical testing.¹⁶ Studies by Qureshi and colleagues¹⁷ and Scalise and colleagues¹⁸ showed similar results after LTO. Biomechanical studies have shown mixed results. Ponce and colleagues¹⁹ showed biomechanically superior results for LTO in comparison to the various repair techniques for ST. In another study, Giuseffi and colleagues²⁰ showed no difference in LTO vs ST during biomechanical testing. In response to the increased concern regarding subscapularis integrity, Caplan and colleagues²¹ reported on 45 arthroplasties in 43 patients with improved postoperative testing with intact subscapularis testing in 90% to 100% of patients. A level 1 randomized control trial conducted by Lapner and colleagues²² did not demonstrate any clear clinical advantage of LTO vs ST. Controversy still exists regarding which is the preferred technique for TSA.

Sanchez-Sotelo and colleagues⁴ evaluated uncemented humeral components in 72 patients who underwent TSA. They found a humeral component was at risk for loosening if a radiolucent line ≥2 mm was present in at least 3 radiographic zones. They also evaluated tilt or subsidence by measurement and whether the components were observed to have changed. Their measured values correlated with their observed values. That study provided a benchmark for evaluation of loosening and subsidence used during this study.⁴ Although radiographic follow-up is limited in this study, we feel that any potential subsidence secondary to use of the LTO technique would be radiographically apparent at 1 year. There were 16 patients without adequate radiographic follow-up included in the study. However, we feel that this was not a large concern, because the study was adequately powered with the patients available to determine a difference based on subsidence.

CONCLUSION

We found no difference in subsidence, lucent lines >2 mm, posterior subluxation, and the Constant and DASH functional outcome scores when we compared TSA performed by a LTO with an ST technique with proximal collar press-fit humeral stem. These data cannot be extrapolated to metaphyseal fit stems, which may exhibit different settling characteristics in the setting of the LTO technique.

This paper will be judged for the Resident Writer’s Award.