Cadaveric Specimens Used in Studies Evaluating Bone Loss in Anterior Shoulder Instability Are Not Representative of the Affected Patient Population: A Systematic Review

Abstract

Purpose

To examine the age, sex, and site of evaluated bone loss in specimens used in cadaveric studies assessing the role of bone loss in anterior shoulder instability and to evaluate whether these studies evaluate glenoid and humeral bone loss separately or as a combined biomechanical problem.

Methods

Embase, MEDLINE, Scopus, Web of Sciences, Google Scholar, and Cochrane databases were queried electronically in April 2023 for cadaveric studies examining the relationship between anatomic structures and recurrent anterior shoulder instability. The review adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses criteria. Studies were excluded if they evaluated the role of soft tissue structures in anterior shoulder stability or assessed anterior bony reconstruction methods to create a stable shoulder and those with fewer than 5 specimens. The following search terms were used: cadaver(s), corpse(s), shoulder dislocation(s), glenohumeral dislocation(s), glenohumeral subluxation(s), glenoid (cavity), and humeral head. Data collected included the age, sex, and site of evaluated bone loss in cadaveric specimens. Studies were categorized based on the anatomic structure that was examined in connection with recurrent anterior shoulder instability: proximal humerus bone defects, glenoid bone defects, and combined bone defects.

Results

Eighteen articles were included in the review, which evaluated 244 cadavers. The sex of 74 cadavers (30.3%) was recorded as male, 50 (20.4%) as female, and 120 (49.1%) were not reported. The mean age of the cadavers was 61.1 ± 15.9 years. Most of the investigations (55.5%) were conducted in the United States. Ten studies looked at the relationship between anterior shoulder instability and glenoid bony structure, 3 assessed the effect of Hill-Sachs lesion size, and only 5 (27.7%) investigated the impact of bipolar bone defects.

Conclusions

The ages of cadavers utilized in biomechanical studies assessing the relationship of glenoid and humeral bone loss to anterior shoulder instability are different from the age range of patients who present with this clinical problem. Males were more frequently studied, and most of these biomechanical studies were conducted in the United States. In addition, few studies evaluate the relationship between anterior shoulder instability and bone loss as a bipolar structural defect.

Clinical Relevance

It is important to understand how closely study specimens match the patient population with the condition being investigated. This study will provide information about the cadaveric specimens used in studies evaluating bone loss in anterior shoulder instability.

The shoulder is the most commonly dislocated joint and is the most likely to suffer recurrent instability after an episode of acute dislocation.1, 2, 3 Shoulder instability is most commonly encountered in adolescents and young adults, with a mean age of anterior capsulolabral fixation of 27.2 years.⁴ Recurrent anterior shoulder dislocation occurs in 10% to 90% of individuals after a primary traumatic dislocation and is more common in young males, collision sports athletes, and military personnel.^2,5, 6, 7 While numerous risk factors for recurrence have been described, both glenoid and humeral bone loss have been studied.^4,8, 9, 10 Glenoid bone loss (GBL) is present in up to 86% of patients with recurrent instability while a Hill-Sachs lesion (HSL) is present anywhere from 73% to 93%. GBL and HSL display interdependence in causing recurrent anterior shoulder instability, and most patients have a bipolar bone defect.11, 12, 13, 14, 15, 16, 17, 18 Therefore, surgeons often incorporate the magnitude of GBL and size of the HSL in their surgical decision-making to determine whether soft tissue and bony procedures, including remplissage and Latarjet, are necessary.19, 20, 21

Several studies have attempted to quantify the amount of critical humeral head^11,22,23 and glenoid bone loss^20,21,24, 25, 26, 27, 28, 29, 30, 31 to predict which patients are at greater risk of recurrence after arthroscopic Bankart repair. A cadaveric study by Yamamoto et al.³¹ demonstrated that glenohumeral stability significantly decreased with glenoid defects greater than 6 mm (20% of the glenoid length and 26% of glenoid width). Although studies have traditionally defined critical glenoid bone loss as 25% of the glenoid width, Shaha et al.³² have suggested that this threshold may be too high in the clinical setting and that glenoid bone loss above 13.5% can lead to suboptimal outcomes after arthroscopic Bankart repair. In another cadaveric study, Kaar et al.¹¹ evaluated glenohumeral stability in the presence of an HSL and showed a decrease in stability with defects greater than 5/8th of the humeral head radius.

Most of the biomechanical studies that evaluate the effect of bone loss on anterior shoulder stability have been conducted using cadaveric specimens. Three widely discussed concerns regarding cadaveric shoulder studies are as follows: glenohumeral capsule integrity (compromised during specimen preparation), scapular orientation, and rotator cuff muscle activation (force is externally applied).³³ Moreover, the cadaveric specimens available for biomechanical studies often belong to people of advanced age whose biomechanical characteristics differ from younger patients in whom this pathology is most common. As a result of increased tissue elasticity in elderly patients (capsulolabral complex) and weakening of the rotator cuff tendons, the mechanism of anterior shoulder dislocation in this population is often due to the disruption of the posterior supporting structures (rotator cuff) rather than failure of anterior capsulolabral complex, often seen in young patients.34, 35, 36, 37 Additionally, differences between male and female cadaveric musculoskeletal structure may influence biomechanical results. These concerns may limit the ability to transfer findings from cadaveric studies directly to the clinical setting.

The purposes of this study were to examine the age, sex, and site of evaluated bone loss in specimens used in cadaveric studies assessing the role of bone loss in anterior shoulder instability and to evaluate whether these studies evaluate glenoid and humeral bone loss separately or as a combined biomechanical problem. We hypothesized that cadaveric studies evaluating bone loss in anterior shoulder instability are performed on a different population than patients with this clinical problem and that many of these studies were not designed to evaluate bone loss on the humeral and glenoid surfaces as bipolar lesions.

Methods

Protocol and Registration

This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines and registered to PROSPERO. A comprehensive search of the literature was performed to retrieve published cadaveric studies evaluating the relationship of glenoid and/or proximal humerus bone loss to recurrent anterior shoulder instability.

Literature Search

The Embase, MEDLINE, Scopus, Web of Sciences, Google Scholar, and Cochrane databases were queried electronically from inception to April 2023. Combinations of the following search terms were utilized: cadaver(s), corpse(s), shoulder dislocation(s), glenohumeral dislocation(s), glenohumeral subluxation(s), glenoid (cavity), and humeral head. No filters were placed on article type, and the references of all relevant articles were carefully checked to ensure that all available articles were considered. Duplicate titles were excluded. The inclusion criteria were as follows: a cadaveric study that was performed to find the relation between anatomic structures and recurrent anterior shoulder instability in all evidence levels without restriction in sex and ages. A language limitation was imposed such that only articles in English were included. Studies were excluded if they evaluated the role of soft tissue structures in anterior shoulder stability or assessed anterior bony reconstructions to create a stable shoulder. Additional exclusion criteria included studies in which the full text of an article was not accessible and those with fewer than 5 specimens (as the findings may be less informative and have a limited impact on clinical practice). Two reviewers (A.K. and A.S-V.) independently screened all titles and abstracts against the inclusion criteria, rejecting those that did not meet all criteria. A second round of full-text screening was subsequently performed for the accepted abstracts. Any disagreements about study exclusion in both abstract and full-text screening were resolved by a third author (R.P.R.).

Data Extraction

Data were collected by 2 independent reviewers, and details of the studies, including country of study, mean age, age range, sex, sample size, anatomic structure under investigation, method of evaluating the influence of bone loss on anterior shoulder stability, and the article’s findings about the impact of bony structure, were documented in an Excel sheet (Table 1). Studies were categorized based on the anatomic structure that was examined in connection to recurrent anterior shoulder instability: proximal humerus bone defects, glenoid bone defects, and combined bone defects.

Table 1.

Article Information and Demographic Data

First Author	Journal	Year of Study	Mean Age, y	Cadavers Number	Sex	Study Nationality
Arciero14	Am J Sports Med	2015	42-65 (range)	21	—	U.S.A.
Bhatia24	J Shoulder Elbow Surg	2019	—	12	12 males	India
Bhatia25	J Shoulder Elbow Surg	2022	—	12	12 males	India
Burns15	Am J Sports Med	2016	73 (57-83)	12	4 males, 8 females	Canada
Gottschalk16	Am J Sports Med	2016	58 (43-69)	18	9 males, 9 females	U.S.A.
Grimberg17	Knee Surg Sports Traumatol Arthrosc	2016	79 (53-99)	7	—	France
Itoi26	J Bone Joint Surg Am	2000	79 (61-99)	10	—	U.S.A., Japan, Korea
Ji27	J Shoulder Elbow Surg	2012	43.21 (21-57)	44	—	Korea
Kaar11	Am J Sports Med	2010	—	8	—	U.S.A.
Lo28	Arthroscopy	2004	74.1 ± 7.4	6	2 males, 4 females	U.S.A.
Nicholson21	J Shoulder Elbow Surg	2022	47-65	8	6 males, 2 females	U.S.A.
Sekiya22	Am J Sports Med	2012	—	9	—	U.S.A.
Sekiya23	Am J Sports Med	2009	—	9	—	U.S.A.
Sang-Jin Shin29	Knee Surg Sports Traumatol Arthrosc	2016	56.5 (43-69)	12	—	U.S.A., Korea
Sang-Jin Shin20	Am J Sports Med	2016	55 (34-68)	8	8 males	U.S.A.
Walia18	Am J Sports Med	2016	57 (43-69)	18	9 males, 9 females	U.S.A.
Wermers30	Knee Surg Sports Traumatol Arthrosc	2021	79.06 ± 8.68 (62-93)	22	7 males, 15 females	Germany
Yamamoto31	Am J Sports Med	2009	74 (68-88)	8	5 males, 3 females	U.S.A.

The primary outcome was the age of cadavers at the time of death, reported as a mean value. Secondary outcomes were country of study, sex of the cadavers, and site of bone defect that was evaluated in relation to anterior shoulder stability (humerus, glenoid, combined) (Table 2).

Table 2.

Article Purpose and Results

First Author	Anatomic Structure Evaluated	Study Purpose	Study Results
Arciero14	Combined	Assess biomechanics of 2 different HSL sizes with increasing amounts of GBL with soft tissue Bankart repair.	Potentiation effect of combined bony glenoid deficiency and HSL. Moderate HSL + glenoid defect of as little as 2 mm → poor outcomes following Bankart repair.
Bhatia24	GBL	1. Quantify changes in 4 glenoid articular parameters in intact and sequential bone loss conditions. 2. Quantify correction after CL and CAL.	CL: consistent correction of most parameters in bone defects of 10% and 20%. CAL: restores and overcorrects articular geometry parameters even in defects of 40%. Articular geometry restorative capability is greater in CAL than CL.
Bhatia25	GBL	1. Quantify critical bone defect size that would significantly alter the bony shoulder stability ratio. 2. Determine restoration of BSSR with Latarjet.	Glenoid cavity depth and BSSR undergo progressive deformation with sequential bone loss. Most of this change happens with 10% glenoid defect. Marginal changes after 10%. BSSR consistently restored in reconstructed glenoid using Latarjet.
Burns15	Combined	Assess interaction of humeral and glenoid bone defects on shoulder engagement. Evaluate sensitivity and specificity of 2 methods.	Support the use of both ABER CT and the IAAA measurement for predicting engagement of bipolar shoulder lesions using 2-dimensional CT.
Gottschalk16	Combined	Define relationship of combine humeral head and glenoid defects.	Decrease in ISR: humeral head defect 44%; glenoid defect 30%; combined humeral head 19 and glenoid defect 20%. Decrease in IT: humeral head ≥31%; glenoid defect ≥20%; combine humeral head 19% and glenoid defect 10%.
Grimberg17	Combined	Biomechanically compare Bankart and remplissage in treatment of Bankart and HSL.	Addition of remplissage procedure to Bankart restores joint stiffness.
Itoi26	GBL	Create and quantify osseus defects and see how effective Bankart repair is at restoring stability and motion.	Osseous size increase → stability decreases without Bankart repair.
Ji27	GBL	Define the most common location of glenoid bony defects in patients with recurrent shoulder dislocation.	Most common location for glenoid bone loss is 3:20 position; more glenoid erosion in patients with anterior instability.
Kaar11	HSL	Investigate size of humeral head defect and resulting glenohumeral instability.	Glenohumeral stability decreases at 5/8 radius defect in ER and abduction. Further instability at neutral and ER when defect is 7/8.
Lo28	GBL	Determine amount of glenoid bone loss required to produce inverted pear glenoid.	The mean amount of bone loss needed for inverted pear shape in patients: 36% (25%-45%); in cadavers: 28.8% (27%-30%).
Nicholson21	GBL	Evaluate biomechanical effects of DAS technique in shoulder with 15% glenoid bone loss in cadavers. Examine DAS with Bankart repair, DAS alone, and Latarjet.	Latarjet resulted in greatest reduction in anterior translation, more than DAS + Bankart, but still Latarjet restrained anterior translation significantly more than native glenoid.
Sekiya22	HSL	Assessing HSL on glenohumeral translations and forces within the glenohumeral capsule and contact of bony surfaces; capsular repair model.	Subtle but significant differences in translations, in situ capsular forces, and bony contact forces between intact and HSLs. Bony forces larger at midranges of motion and capsular forces larger at extremes of motion.
Sekiya23	HSL	Determine effects of HSL regarding size, shoulder position, and osteoarticular repair on glenohumeral translation and stability in response to compressive and anterior load.	Osteoarticular repair that restored glenohumeral stability defects as small as 12.5% can have biomechanical consequences.
Sang-Jin Shin29	GBL	Determine relationship between bony stability and percentage of anterior glenoid bone loss and determine effect of bone loss on orientation.	As glenoid defect size increased, dislocation force required decreased proportionally. Any amount of bone loss from anterior margin on glenoid results in significant loss of glenohumeral bony stability.
Sang-Jin Shin20	GBL	Find critical value of anterior glenoid bone loss when a soft tissue repair does not restore rotational range of motion, anterior-inferior glenohumeral translation, and normal humeral position using biomechanical anterior shoulder instability model.	Decreased ER, increased anteroinferior translation with external load, and increased posterior and inferior shift of humeral head apex in position of maximum ER after soft tissue Bankart repair of 15% glenoid defect compared with repair of Bankart without glenoid defect. Fifteen percent glenoid defects or more should be considered critical bone loss amount at which soft tissue repair cannot restore glenohumeral translation.
Walia18	Combined	Evaluate influence of combined bony lesions on shoulder instability through varying glenohumeral positions.	Humeral head defect → rotational instability; significant glenoid defect → loss of translation independent changes. Combined defects as small as 19% humeral head and 10% glenoid led to significant instability.
Wermers30	GBL	Evaluate whether glenoid concavity allows a more precise assessment of glenohumeral stability than the defect size alone.	Glenoid concavity is a significant predictor for SR. Defect size provides minimal informative value.
Yamamoto31	GBL	Would the anterior glenoid defect have a similar effect on anterior shoulder stability to the anteroinferior glenoid defect? Determine the critical size of the anterior glenoid defect created at 3 o’clock on anterior stability of the shoulder.	An osseous defect at 3 o’clock with a width that was equal to or greater than 20% of the glenoid length significantly decreased anterior stability.

ABER, abduction and external rotation; BSSR, bony shoulder stability ratio; CAL, congruent-arc Latarjet; CL, classic Latarjet; CT, computed tomography; DAS, dynamic anterior shoulder stabilization; ER, external rotation; GBL, glenoid bone loss; HSL, Hill-Sachs lesion; IAAA, intact anterior articular angle; ISR, intact stability ratio; IT, intact translation; SR, stability ratio.

Statistical Analysis

Descriptive analysis was used to examine information on the number of cadavers, the specimens’ demographics, the country where the study was conducted, and the anatomic structure that was assessed. Pooled mean age was calculated along with standard deviation. When the standard deviation was not reported in the original study, it was estimated by the method of Hozo et al.³⁸ using the following formula: estimated standard deviation = [“upper end of range” – “lower end of range”] / 4.

Results

Literature Search

Search of Embase, MEDLINE, Scopus, Web of Sciences, Google Scholar, and Cochrane databases found 1,792 results, of which 870 were duplicates and thus removed. A total of 681 studies were excluded by assessing whether the abstracts met the inclusion and exclusion criteria. The full-text manuscripts of the remaining 241 potentially relevant articles were scrutinized. Of these, 148 studies were eliminated for evaluating factors other than bony anatomy on anterior shoulder stability. Of the 93 articles left, only 22 precisely evaluated the relationship between the bony defect on either side of the joint and anterior shoulder stability, and only 18 had full texts that could be accessed in English and evaluated more than 5 cadaver specimens. A Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart summarizes the results of the literature search and screening process (Fig 1).

Fig 1.

The systematic review Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart.

Study Population

A total of 244 cadavers were incorporated in the systematic review. Of the 244 cadavers, 74 (30.3%) were recorded as male, 50 (20.4%) as female, and 120 (49.1%) were not reported. Seven studies (38.8%) did not disclose the mean age of the cadavers that were investigated or only reported the age range, whereas 11 studies recorded the mean age of their specimens. The mean age of the cadavers that were studied for the effect of bony anatomy on anterior shoulder instability was 61.1 $\pm$ 13.8 (21-99) years. The geographical distribution of studies was as follows: 10 (55.5%) studies were done entirely in the United States, while 6 (33.3%) were performed outside of the United States (India, South Korea, Canada, France, and Germany), and 2 (11.1%) were completed in multiple countries. Ten (55.5%) studies only analyzed the relationship between glenoid bone loss and anterior shoulder instability, 3 (16.6%) only assessed the effect of HSL size on shoulder instability, and 5 (27.7%) investigated the impact of combined bone deficiency on anterior shoulder instability.

Discussion

The main finding of this systematic review is that the mean age of cadavers used to evaluate the relationship between glenoid/humeral bone loss and recurrent anterior shoulder instability is significantly older than that of the patients who present with this clinical problem. Although most of these biomechanical cadaveric studies did not disclose their specimen sex, males were more thoroughly investigated than females, and most of the studies were conducted in the United States. Few studies addressed the relationship between combined bone defects and anterior shoulder instability.

Although anterior shoulder instability is a well-known disorder of the younger population, as reported broadly in the literature by Verweij et al.⁴ showing that the mean age of patients who underwent anterior capsulolabral shoulder fixation due to instability was 27.2 years, the biomechanical cadaveric studies in this systematic review were performed in a significantly older age group (61.1 years). These findings are particularly noteworthy, as the paradigm of critical glenoid bone loss and the glenoid track concept are largely attributable to landmark cadaveric studies, including those by Itoi et al.²⁶ and Yamamoto et al.,³¹ who demonstrated that glenoid bone loss greater than the critical threshold values of 21% and 20%, respectively, present with a greater risk of recurrent anterior shoulder instability. However, both of these studies utilized cadavers with a mean age well above 70 years. Although clinical studies have since corroborated these concepts, it is important to note that these specific thresholds were derived from cadaveric specimens dissimilar to the clinical population treated with this condition. This could potentially serve as a reason to why critical glenoid bone loss thresholds have since been refined in recent clinical studies, with some reporting values as low as 13.5%.³²

Patient age has a direct effect on the size of the various anatomic structures: many structures are predicted to be smaller in children and adolescents than in adults. Furthermore, it is well recognized that aging reduces skeletal muscle mass and cartilage thickness, causing changes in bone shape and geometry as well as tendon degeneration.39, 40, 41 Volpi et al.⁴¹ showed a 3% to 8% reduction in muscle mass per decade from age 30 and further deterioration after age 65, which would reduce the efficacy of shoulder muscles as a dynamic stabilizer. Specifically addressing the glenohumeral joint, Knapik et al.,⁴² after evaluating 993 adult specimens, reported a 0.056-mm increase in glenoid width, a 0.043-mm increase in coracoid width, a 0.011-mm increase in coracoid height, and a 0.044-mm increase in coracoid length per each year of aging, which may become clinically significant while taking decades into account. In situations where surgical decision-making relies on anatomic measurements on the scale of millimeters, as in patients with both recurrent anterior shoulder instability and bony defects, the difference in age of patient populations as well as the cadavers referenced for decision-making can likely affect the expected surgical outcome. As previously mentioned, there are additional concerns to consider when basing a treatment approach on cadaveric studies, which Williamson et al.³³ highlighted. These concerns include increased tissue elasticity in elderly patients, scapular orientation, rotator cuff muscle activation, and glenohumeral capsule integrity.

In this review, we found that 4 (22%) of the 18 included articles did not report the ages of their samples. This may have been a result of cadaver age being unknown, a common limitation due to unclaimed cadavers that are likely to lack age information. In fact, Habicht et al.,⁴³ in their evaluation of the source of cadavers used for education in medical schools in 68 countries, found that 31% utilized unclaimed cadavers while 37% used a mixture of unclaimed and donated bodies.

While substantial anatomic structures and size differences between genders have been anticipated,^42,44, 45, 46, 47 Knapik et al.⁴² showed a significantly wider glenoid in their male cadaver species. Similarly, Merrill et al.⁴⁸ evaluated gender differences in the size and overall shape of the glenoid by measuring glenoid height and width, notch location, and depth in 363 cases. They found significant differences between female and male anatomy for each dimension measured. In this review article, like in many other cadaveric studies, more male cadavers were studied than female (30.3% vs 20.4%). This highlights significant limitations using cadaveric data to guide surgical decisions for female patients with recurrent anterior shoulder instability. Race is another major concern in anatomic studies. Larger mean glenoid widths have been shown in Caucasians in comparison to African Americans.⁴² Most cadaveric investigations on recurrent anterior shoulder instability have been conducted in the United States (55.5%). Thus, extra caution should be used when applying cadaveric-derived anatomic measures in surgical decision-making for patients of different nationalities. However, it is also important to recognize that anatomic cadaveric studies depend on the availability of bodies, which in turn depends on many factors, like national legislation, cultural and religious customs, and the willingness of the population to contribute to anatomic education.^43,49

According to this systematic review, only a few biomechanical cadaveric studies investigating the influence of bony defects on anterior shoulder stability chose to study both GBL and HSL in combination (27.7%). Most of these studies (55.5%) assessed only glenoid bone structure on glenohumeral joint stability without considering the influence of a concomitant HSL even though proximal humerus bone lesions are seen in 70% of first-time dislocations and almost all patients with recurrent shoulder instability.50, 51, 52 GBL and HSL display interdependence in causing recurrent anterior shoulder instability, and most patients in the clinical setting have a bipolar bone defect.11, 12, 13^,53, 54, 55 Thus, it is important to rely on studies that emphasize the bipolar nature of bone loss in anterior shoulder instability.

The ages of the cadavers studied for evaluating the relation of bone loss and anterior shoulder instability differ significantly from those of patients with this clinical issue. Most of the cadavers used in these studies are male, and the majority of the research was conducted in the United States; exercising caution is advisable when extrapolating the information from these articles directly for surgical decision-making, particularly for populations that are not well represented in these studies, including females and patients outside the United States.

Limitations

This study is not without limitations. This study is subject to the limitations of most systematic reviews. First, there is a possibility of search bias due to the difficulty of screening for every possible study assessing the subject of interest. Second, a selection bias element was likely present in most of the investigations, which were controlled laboratory studies with small samples derived from the available resources. Third, the statistical analysis is limited because not all the included publications reported information on cadaveric age and gender. This highlights a prevalent limitation of many cadaveric studies, as Yammine,⁴⁹ in a study of 390 cadaveric studies, found that 28% did not report the age of their samples. Furthermore, there was considerable heterogeneity among the studies in their method of biomechanical testing, bony defect creation, and soft tissue handling that precluded the ability to evaluate the critical size of bony defect for recurrent anterior shoulder stability based on the existing cadaveric studies assessed in this systematic review. Although there is a logical inference based on anatomic studies, there is also limited evidence that cadaveric age and sex directly correlate to biomechanical outcomes. The heterogeneity of study purposes and methods did not allow for a comparison of biomechanical outcomes between studies based on age or gender.

Conclusions

The ages of cadavers utilized in biomechanical studies assessing the relationship of glenoid and humeral bone loss to anterior shoulder instability are different from the age range of patients that present with this clinical problem. Males were more frequently studied, and most of these biomechanical studies were conducted in the United States. In addition, few studies evaluate the relationship between anterior shoulder instability and bone loss as a bipolar structural defect.

Disclosures

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: A.L. is a paid consultant for Arthrex and Stryker not related to this study. All other authors (A.K., A.S-V., R.P.R., M.P.K.) declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.