**Prevalence of Meibomian Gland Dysfunction and Its E**ff**ect on Quality of Life and Ocular Discomfort in Patients with Prosthetic Eyes**

#### **Alessandro Meduri 1,\*, Rino Frisina 2, Miguel Rechichi <sup>3</sup> and Giovanni William Oliverio <sup>1</sup>**


Received: 23 April 2020; Accepted: 8 June 2020; Published: 9 June 2020

**Abstract:** Purpose: To evaluate the influence of ocular discomfort and meibomian gland dysfunction (MGD) on quality of life in patients with an ocular prosthesis. Methods: a prospective analysis was conducted on 18 patients with a unilateral ocular prosthesis. Evaluation of ocular discomfort symptoms, lid margin abnormalities (LMA), meibomian gland expression, meibography and a psychometric evaluation using the National Eye Institute Visual Function Questionnaire (NEI VFQ), Facial Appearance subscale of the Negative Physical Self Scale (NPSS-F), Hospital Anxiety and Depression Scale (HADS) and the DAS24 to evaluate anxiety and depression. Results: the statistically significant differences observed between normal and prosthetic eyes related to ocular symptoms and the meibography score (*p* = 0.0003). A negative correlation was reported between NEI VFQ score and meibography score (r = −0.509; *p*-value = 0.022). A positive correlation was detected with NPSS (r = 0.75; *p*-value < 0.0001), anxiety HADS score (r = 0.912; *p*-value = 0.001) and depression HADS score (r = 0.870; *p*-value > 0.0001). Conclusion: MGD represents the most common cause of evaporative dry eye disease, due to the reduction of the thickness of the lipid layer of the tear film. The occurrence of MGD in patients with prosthetic eyes is very common. Anxiety and depression were correlated to ocular discomfort and MGD, and this could affect the quality of life in patients with an ocular prosthesis.

**Keywords:** ocular prosthesis; meibomian gland dysfunction; quality of life; ocular discomfort

#### **1. Introduction**

The loss of an eye presents severe emotional stress and commonly results in negative consequences to social behavior. Patients with an ocular prosthesis reported a high level of anxiety and depression as a consequence of their changed appearance. In this sense, the use of an ocular prosthesis purposes to improve cosmetic appearance, as well as ameliorating social acceptance [1]. Nowadays, ocular prosthesis can be stock or patient-customized, considering several factors, such as the anatomy and tissue bed of the socket, and individualized aesthetic requirements. The main material used for the fabrication of ocular prostheses is polymethyl methacrylate (PMMA), which is compatible with tissues, and has easy color modification abilities and elevated aesthetic appearance. Numerous techniques to improve the aesthetic results of the ocular prosthesis have been proposed, such as painting the sclera and iris, the use of transparent grids for proper orientation and the use of a digital image of the normal eye [2]. The principal causes for the need of an ocular prosthesis comprise tumors, congenital defects and malformations, irreparable trauma, end-stage eye diseases, and severe ocular diseases associated with uncontrolled pain, such as neovascular glaucoma, or an unattractive appearance, such

as phthisis bulbi [3,4]. Eye removal surgery is classified into evisceration, enucleation and exenteration. Evisceration includes the removal of the ocular contents, leaving in place the sclera shell. The mobility of the eviscerated globe implant is preserved, since the extraocular muscles are intact. Enucleation consists in the removal of the eyeball after the extraocular muscles and the optic nerve have been separated. Sufficient space is formed for the prosthesis, with movement of the fornix within the enucleated socket providing mobility to the ocular prosthesis. Exenteration is the surgical removal of the complete contents of the orbit. The eyelids may or may not be involved. Exenteration defects in some instances may be allowed to heal by secondary intent, but adequate space must remain in the resultant defect to allow the prosthesis to be positioned superiorly and posteriorly enough for a good cosmetic appearance [5]. An orbital implant is used to fill the orbital cavity or the scleral shell, and to replace the consequent volume reduction due to the removal of the eye. The eye's motility is preserved, as the scleral implant is attached to the ocular muscles. Several types of implants exist, differing in shape (spherical or oval), being stock or customized, porous or non-porous, and in the presence of a peg or motility post [4,5]. Chronic ocular discomfort represents one of the most common adverse events referred to in long-standing patients with an ocular prosthesis. Many patients with an ocular prosthesis report varying degrees of ocular discomfort, such as discharge, dryness, irritation and a sticky sensation. Numerous mechanisms that could lead to ocular discomfort in prosthetic eye wearers were suggested, such as the infection of the anophthalmic socket, glutinous surface deposits and a roughened prosthesis [6,7]. A recent study analyzed the role of the tear film and demonstrated the linear correlation between tear deficiency and discomfort reported in patients with an ocular prosthesis [7]. Recent studies also confirmed the relationship between MGD and contact lens wear, as well as in ocular prosthesis wear, demonstrating a larger grade of Meibomian gland loss and alterations compared with normal paired eyelids [8,9]. Meibomian gland dysfunction (MGD) represents the most common cause of evaporative dry eye disease, due to the reduction of the thickness of the lipid layer of the tear film [10]. The aim of this study is to evaluate the prevalence of MGD in patients with ocular prostheses reporting symptoms of ocular discomfort, and their impact on quality of life.

#### **2. Results**

Demographic data is reported in Table 1. Eighteen patients of mean age 51.3 ± 16.8 years and mean length of prosthesis duration 13.6 ± 5.1 years were studied. The most frequent causes of blindness were trauma (ten patients), cancer (five patients) and other ocular diseases (three patients) such as endophthalmitis and one case of neo-vascular glaucoma. A case of unilateral ocular prosthesis due to neo-vascular glaucoma is reported in Figure 1. Evisceration was the most common surgery (10 patients). All ocular prostheses were polymethyl methacrylate (PMMA). Sixteen patients had a customized prosthesis, presenting a good aesthetic appearance. Only two patients had a stock prosthesis. In Table 2 summarises scores regarding ocular symptoms, lid margin abnormalities (LMA), meibomian gland expression and meiboscore related to the normal eye and the prosthetic eye in the upper and lower eyelid. Significant statistical differences were noted between the normal eye and prosthetic eye related to ocular symptoms (*p*-value < 0.0001), LMA (*p* = 0.0006), meibomian gland expression score (*p* = 0.0003), and in the meiboscore of the upper eyelid (*p* = 0.0004), lower eyelid (*p* = 0.0003) and total meiboscore (*p* = 0.0003). A positive correlation between duration of prosthesis and meiboscore in the prosthetic eye r = 0.878; *p* < 0.001 95% CI: [0.688, 0.956] was observed. All psychometric data are reported in Table 3. A negative correlation was reported between NEI VFQ score and meiboscore (r = −0.509; *p*-value = 0.022). Positive correlations were detected with NPSS (r = 0.75; *p*-value < 0.0001); anxiety HADS score (r = 0.912; *p*-value < 0.001); depression HADS score (r = 0.870; *p*-value < 0.0001); DES24 (r = 0.686; *p*-value < 0.0001) and VAS score for sadness (r = 0.657; *p*-value = 0.002) anger (r = 0.741; *p*-value < 0.0001) and shamefulness (r = 0.744; *p*-value < 0.0001). (Figure 2 and Table 4).


**Table 1.** Clinical Characteristics of the Study Population.

Data are presented as n (%), and mean ± standard deviation.

**Figure 1.** Ocular prosthesis (**A**), anophthalmic socket (**B**).

**Table 2.** Comparison of meibomian gland and lid margin evaluation in prosthesis and normal eye.


All data are reported as mean ± standard deviation. \* Wilcoxon signed-rank test. Bold characters for *p*-value < 0.05.



Vision-specific composite refers to NEI VFQ score; appearance concern refers to NPSS score; shame, sadness and anger were evaluated using a visual analogue score. All data are reported as mean ± standard deviation.

*Prosthesis* **2020**, *2*

**Figure 2.** Scatter plot showing the relationship between total meiboscore (total) and (**A**) Hospital Anxiety and Depression Scale (anxiety score); (**B**) Hospital Anxiety and Depression Scale (depression score); (**C**) Negative Physical Self Scale-Facial appearance subscale score; (**D**) National Eye Institute Visual Function Questionnaire; (**E**) Derriford Appearance Scale; (**F**) Shameful (visual analogue scale); (**G**) Sadness (visual analogue scale); (**H**) Anger (visual analogue scale).


**Table 4.** Pearson's correlation between Meiboscore (total) and psychometric score.

Legend: NEI VFQ—The National Eye Institute Visual Function Questionnaire; NPSS-Negative Physical Self Scale; HADS—Hospital Anxiety and Depression Scale; DAS-24—Derriford Appearance Scale.\* ANOVA test, bold characters for *p*-value < 0.05.

#### **3. Discussion**

Chronic ocular discomfort represents one of the most common adverse events referred to in long-standing patients with an ocular prosthesis. Chronic discharge, in patients with an ocular prosthesis, could be related to several conditions. In particular, they could be classified as prosthesis-related, such as poor fit, mechanical irritation, reaction to deposits and poor prosthesis hygiene; socket and implant-related, such as exposure, granuloma, peg-related, socket contraction and environmental allergens; lacrimal-related, such as reduced tear production or outflow obstruction; or eyelid-related, such as MGD, lagophthalmos and lack of mucous membrane or skin. A common condition of ocular discomfort in a long-standing patient with ocular prosthesis is giant papillary conjunctivitis, due to a prolonged mechanical irritation and the immunologic reaction of the ocular surface [11,12]. Rokohl et al. documented a relationship between the discharge severity of prosthetic eyewear with conjunctival inflammation, higher cleaning frequency and less hand washing before handling [13]. The role of reduced tear production was documented in previous studies. A reduced tear volume was documented in anophthalmic patients, as demonstrated by a reduced Schirmer's test value, as well as a low tear meniscus height [14,15].

Recent studies have also considered the morphologic changes in meibomian glands related to prosthetic eye patients, demonstrating a larger grade of meibomian gland loss compared with the normal paired eyelids [8,9]. In the present study, a high prevalence of MGD was recognized in the population of patients evaluated. In particular, significant differences were demonstrated between the normal eye and the prosthetic eye, concerning ocular symptoms (*p*-value > 0.0001), LMA (*p* = 0.0006), meibomian gland expression score (*p* = 0.0003), lid margin abnormalities (*p* = 0.0006) and meiboscore (*p* = 0.0003). Meiboscore was evaluated using Keratograph 5M (Oculus, Wetzlar, Germany) to acquire a noninvasive meibography. MGD describes a group of disorders characterized by functional abnormalities of the meibomian glands. The International Workshop defined MGD a chronic, diffuse abnormality of the meibomian glands, commonly characterized by terminal duct obstruction and/or qualitative/quantitative changes in glandular secretion [16]. MGD can lead to altered tear film composition, ocular surface disease, ocular and eyelid discomfort, and evaporative dry eye. In fact, MGD represents the principal cause of evaporative dry eye [17]. Eyelids of a prosthetic eye seem mostly disposed to obstructive MGD, as result of an increased hyperkeratinization, causing excretory duct obstruction, due to a combination of tear deficiency, deposit accumulation, and micro-trauma. Other hypothetical pathophysiological mechanisms of meibomian gland obstruction, in the eyelid of patients with an ocular prosthesis, is decreased and weakened eyelid blinking. Recently, it was suggested that tear insufficiency as consequent of MGD may be the cause of ocular discomfort for those patients, particularly in whom no specific etiology can be identified. Dry eye symptoms are common in long-standing ocular prosthesis wearers. In this study, a significant difference was demonstrated by comparing the normal eye and prosthetic eye (*p*-value < 0.0001) relative to ocular discomfort symptoms. Duration of the prosthesis may play a central role; thus, a positive correlation

with meiboscore in prosthetic eye was seen (r = 0.878, *p* < 0.001). Jang, et al. reported a greater degree of meibomian gland loss in patients who had used an ocular prosthesis longer than 10 years. In our group of patients, the average duration of prosthesis use was 13.6 ± 5.1 years. For this reason, MGD could represent an undiagnosed cause of chronic ocular discomfort in long-standing ocular prosthesis patients. The occurrence of biofilm growth has been confirmed on many ocular prostheses. Biofilms may play a significant role in the tolerability of the ocular prosthesis. However, biofilm has also been identified on the surface of scleral implants, without signs of clinical infection. The longstanding ocular prosthesis has been related to the development of giant papillary conjunctivitis, resulting in poor tolerability [18]. Litwin, et al. analyzed the role of the degree of prosthesis surface polishing, comparing the standard and the optical polishing procedures. They reported better tolerability at 12 months in the optical polish of the prosthesis group [19]. Patients with an ocular prosthesis reported a high level of social anxiety, and avoid social situations as a consequence of their altered appearance. This was documented by the mean score (2.8 ± 1.13) relative to the Facial Appearance subscale of the Negative Physical Self Scale (NPSS). A score > 2 was observed in eight patients, indicating displeasure with facial appearance. Similar results were also noted relative to the DAS24 questionnaire score. The mean score of DAS24 in our group of patients was 45.88 ± 25.63.

DAS24 is a measure of social anxiety and social avoidance in relation to appearance. Sixteen patients had a customized prosthesis, presenting a good aesthetic appearance.

The lowest questionnaire scores were found only in the two patients with stock prostheses.

The HADS is a psychometrical scale assessing patients' anxiety and depression levels related to health problems. Considering the score of the HADS questionnaire, a high level of depression (6.29 ± 2.77) and anxiety (6.76 ± 2.33) were detected in our population of patients.

Chronic ocular discomfort could represent a factor that increases the anxiety condition, in patients with an altered appearance. Several studies have shown an association of dry eye disease with depression and anxiety. However, a direct correlation between the severity of dry eye disease and anxiety or depression was not demonstrated [20]. It is plausible to consider an increased risk of developing anxious and depressive disorders in patients with an ocular prosthesis with symptoms of dry eye and MGD. In fact, a higher score of anxiety and depression was recognized in patients presenting symptoms and signs of MGD. In particular, a statistically significant direct correlation was observed between meiboscore and the anxiety HADS score (r = 0.912; *p*-value = 0.001), as well as with the depression HADS score (r = 0.870; *p*-value > 0.0001 (Figure 2)).

Nevertheless, several factors could influence the psychological behavior of patients, and a chronic ocular discomfort could represent a single factor inside a wide variety of causes. In conclusion, this study suggests that anxiety and depression conditions could be common in patients with eye prostheses, in particular in the presence of chronic ocular discomfort caused by MGD. This result should be confirmed in a bigger cohort of patients, to find other factors, such as sex, age, therapy and metabolic differences, that could play a direct influence.

#### **4. Materials and Methods**

A prospective, multi-centric analysis was conducted on 18 patients (ten male, eight female) with a unilateral ocular prosthesis, between July 2018 and January 2020. The study was conducted in accordance with the tenets of the Declaration of Helsinki. All patients signed an informed consent after a full explanation of all study-related procedures. Inclusion criteria were ability and willingness to participate in the study, unilateral ocular prosthesis worn at least 2 years and subjective symptoms and signs of ocular discomfort such as burning, wetness, foreign body sensation, pain, itching, dryness, tearing and mucous discharge. Patients with inflammation or infection of the socket, or with a poorly fitted prosthesis were excluded from the study. Each patient was questioned about their duration of prosthesis use and the type of ophthalmic surgery (such as evisceration or enucleation) requiring ocular prosthesis.

#### *4.1. Symptom Evaluation*

Patients were asked to assign severity using this scale for the following symptoms: burning, wetness, foreign body sensation, pain, itching, dryness, tearing, mucous discharge, hyperemia, excessive blinking, uncomfortable in windy conditions and uncomfortable in dry conditions, evaluated according to a scoring system from 0 (absent) to 3 (severe). A global score was obtained by summing up the scores of each symptom and the values (score range 0–36). The assessment of symptoms was the same for the prosthetic eye-wearing side and for the normal side.

#### *4.2. Meibomian Gland and Lid Margin Evaluation*

A slit-lamp examination was conducted to evaluate lid margin abnormalities and meibomian gland expression. Lid margin abnormalities were scored from 0 to 4 based on the presence or absence of the following parameters: irregular lid margin, plugging of meibomian gland orifices, vascular engorgement, or a shift in the mucocutaneous junction [10]. Meibomian gland expression was assigned grades for clarity and ease of meibum expression (grade 1–4). Meibography was performed using a Keratograph 5M (Oculus, Wetzlar, Germany) in the upper and lower eyelid separately. Meiboscore grading was assessed by Keratograph 5M automatic software (JENVIS Meibo Grading Scales). Grade 0: no loss of Meibomian glands; Grade 1: loss of less than 1/3 of the total meibomian gland area; Grade 2: loss of 1/3 to 2/3 of the total area; Grade 3: loss of more than 2/3 of the area. The total Meiboscore was considered as the sum of the meiboscore of the upper eyelid and lower eyelid [21].

#### *4.3. Psychosocial Variables*

Four standardized questionnaires assessing quality of life, anxiety and depression related to physical health problems, social anxiety and social avoidance in relation to appearance, and three supplementary elements examining the level of shamefulness, sadness and anger felt using a visual analogue scale (0–10). The National Eye Institute Visual Function Questionnaire (NEI VFQ) is a questionnaire assessing the quality of life related to visual function, comprising several subcategories including general and peripheral vision, color perception, ocular discomfort, near and distance activities, social performance, mental healthiness and driving ability. A complex score is estimated by the average of the subcategories' scores (0–100). Lower scores denote poor functioning [22]. The Facial Appearance subscale of the Negative Physical Self Scale (NPSS) is an effective measure of appearance, comprising five categories. We considered the Facial Appearance subscale. The NPSS-F score is estimated by the average of the item scores (>2 indicating displeasure with facial appearance) [23]. The Hospital Anxiety and Depression Scale (HADS) was used to evaluate the level of anxiety and depression. The HADS comprises seven items (scored 0–3) about anxiety, and seven items about depression (scored 0–3). Total scores range from 0 to 21 for both subcategories, with lower scores signifying low levels of anxiety or depression [24]. The DAS24 is a measure of social anxiety and social avoidance in relation to appearance. The total score ranges from 11 to 96, with high scores signifying high levels of social anxiety and social avoidance [25].

#### *4.4. Statistical Analysis*

The numerical data were expressed as mean and standard deviation, and the categorical variables as absolute frequency and percentage. Examined variables did not present normal distribution, as verified by the Kolmogorov Smirnov test; consequently, the non-parametric approach was used. For each parameter, we performed statistical comparisons between the normal eye and the prosthetic eye in the exam, using the Wilcoxon signed-rank test for numerical variables. The Pearson correlation coefficient (r) was calculated to measure the strength of correlations between parameters. Statistical analyses were performed using JASP (Version 0.11.1). A *p*-value smaller than 0.05 was considered to be statistically significant.

**Author Contributions:** Conceptualization, G.W.O. and A.M.; methodology, G.W.O.; software, G.W.O.; validation, A.M., R.F. and M.R.; formal analysis, G.W.O.; investigation, A.M.; resources, A.M.; data curation, G.W.O.; writing—original draft preparation, G.W.O.; writing—review and editing, A.M.; visualization, R.M.; supervision, A.M.; project administration, A.M.; funding acquisition, A.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **A Finite Element Model for Trigger Finger**

#### **Helena I. Relf 1, Carla G. Barberio <sup>2</sup> and Daniel M. Espino 1,\***


Received: 12 June 2020; Accepted: 20 July 2020; Published: 22 July 2020

**Abstract:** The aim of this study was to develop a finite element model to investigate the forces on tendons which ensue due to trigger finger. The model was used to simulate both flexor and extensor tendons within the index finger; two test cases were defined, simulating a "mildly" and "severely" affected tendon by applying constraints. The finger was simulated in three different directions: extension, abduction and hyper-extension. There was increased tension during hyper-extension, with tension in the mildly affected tendon increasing from 1.54 to 2.67 N. Furthermore, there was a consistent relationship between force and displacement, with a substantial change in the gradient of the force when the constraints of the condition were applied for all movements. The intention of this study is that the simulation framework is used to enable the in silico development of novel prosthetic devices to aid with treatment of trigger finger, given that, currently, the non-surgical first line of treatment is a splint.

**Keywords:** biomechanics; finite element; hyper-extension; trigger finger; stenosing tenosynovitis; tendon; tension

#### **1. Introduction**

The hand is considered the most dexterous and well-coordinated part of the body, having great complexity and utility [1,2]. Its mobility is vital for any individual's independence during daily activities. Stenosing tenosynovitis, more commonly known as trigger finger, is one of the most common pathologies seen in hand surgery [3–5]. In a healthy hand, the flexor tendon should be able to move freely inside the tendon sheath. However, in this condition, the tendon and/or sheath become inflamed or irritated, forming scar tissue due to fibrocartilagenous metaplasia of the tendon, which restricts tendon movement through the sheath [6,7].

This restriction from trigger finger can result in painful locking and clicking of the finger [3,5,8–10]. "Triggering" refers to the sudden release of the tendon after catching during finger extension. Trigger finger is more commonly found in healthy middle-aged women [4,8,10,11] but is also associated with conditions such as diabetes, arthritis [10–13] and carpal tunnel syndrome [5,13]. The exact cause of the condition is unclear and can vary between cases, but the tendon can be further aggravated by hand use at work or during sport [12]. Symptoms include tenderness in the affected area, movement pain and locking or clicking; if attended to promptly, pain and swelling can be reduced easily. Surgery is only considered if other treatment options fail or the condition goes untreated for an extended period of time [8]. Non-surgical treatments include splinting, physiotherapy, nonsteroidal anti-inflammatory drugs and corticosteroid injections [3,8,10,12]. Surgery is performed under local anaesthetic, with an incision created in the roof of the tendon sheath in order to widen the tunnel so that the tendon can move freely [14].

The hand is comprised of twenty-seven different bones consisting of phalanges, metacarpals and carpals (Figure 1). Effective function is coordinated by a linkage system of tendons, ligaments and muscles. Tendons transmit loads from the bones to the intrinsic muscles and are interconnected by aponeuroses; this is commonly referred to as the extensor mechanism [15,16]. All fingers have an extensor tendon, located on the posterior surface of the hand, and two flexor tendons, located on the palmar side; furthermore, the second and fifth fingers have an additional extensor tendon. The extensor digitorum communis (EDC) straightens the finger [17] from both the proximal and distal interphalangeal joints. Flexing (bending) of the finger is achieved through the flexor digitorum profundus (FDP) and the flexor digitorum superficialis (FDS) tendons that connect to the distal and middle phalanx, respectively. Flexor tendons are channelled through and constricted by the tendon sheath [18], with some lubrication provided by synovial fluid.

**Figure 1.** Schematic diagram of the hand. (**a**) Arrangement of the bones in the hand. (**b**) Simplified diagram of the attachment point of tendons in a finger.

There is limited research on the biomechanical impact of trigger finger and how it immobilises and generates stress within the hand. Most existing studies have been based on physical examinations of patients with the condition, using methods such as motion analysis and electromagnetic tracking systems. The higher the grade, the lower the range of movement for each joint in the finger, and restricted tendon mobility can measurably change comparative exerted force between the thumb and fingers [19,20]. Long-term, however, even after treatment, some disability may persist [21]. While a few biomechanical models are available for determining forces in tendons of the hand, limitations persist for their extrapolation to an understanding of the mechanics of trigger finger. For example, some models are specific to climbing techniques [22]. Others explore the finger extensor tendon network but do not model trigger finger [2]. Arguably, the most comprehensive model is provided by Lu et al. [23] to evaluate the forces within different tendons, predicting higher tension in the FDP tendon than in the FDS tendon when both tendons were triggering. There is scope, though, to evaluate the effect of trigger finger on the tension in the tendons of the hand during a range of movements. Such a model would enable a simulation framework to be available to test future prosthetic devices intended to aid in trigger finger "treatment", as an objective technique for early stage development.

The aim of this study is to develop a model, using finite element analysis (FEA), which can predict the forces within tendons in the hand during trigger finger. The focus is on the right hand's index finger as this is where trigger finger is most likely to occur [14]. Models developed include a healthy case along with both a mild and a severe model for trigger finger to enable direct comparison. The contribution made by this study is, therefore, in developing an FEA model for mild and severe levels of stenosing tenosynovitis (i.e., trigger finger).

#### **2. Results**

#### *2.1. Outline of Results*

Figures 2–8 present the results for the forces acting on the tendons and their displacement. Stress–strain curves have also been plotted to analyse the data. Forces on the tendons followed a nonlinear relationship, with clear differences between healthy tendons and those with the constraints of mild and severe trigger finger. The effects of extension, abduction and hyper-extension are outlined in Sections 2.2–2.4, with Section 2.5 outlining the variation of tension and cross-sectional area within a tendon.

**Figure 2.** Tendon displacement for position A for healthy, mildly and severely affected tendons.

#### *Prosthesis* **2020**, *2*

**Figure 3.** Force predicted for flexor digitorum profundus (FDP) and flexor digitorum superficialis (FDS) tendons (position A). (**a**) Force against time for the FDP and FDS tendons, (**b**) force against displacement of the FDP and FDS tendons.

**Figure 4.** Force predicted for the extensor digitorum communis (EDC) tendon (position A). (**a**) Force against time, (**b**) force against displacement.

**Figure 5.** Time-dependent displacement of the FDS tendon (position A; note: TF: trigger finger).

*Prosthesis* **2020**, *2*

**Figure 6.** Force predicted for the FDP and FDS tendons (position B). (**a**) Force against time, (**b**) force against displacement.

**Figure 7.** Force predicted for the FDP and FDS tendons (position C). (**a**) Force against time, (**b**) force against displacement.

**Figure 8.** Stress–strain predicted for the FDP and FDS tendons (position C).

#### *2.2. Position A—Extension*

The locking of the tendon, caused by trigger finger, had a notable effect on the forces exerted in the tendons. In the case of severe trigger finger, a sharp increase in exerted force is observed at the point at which the tendon is restricted, in contrast to the healthy tendon where force increases gradually over time (Figures 2 and 3). There are much greater forces on the severely affected tendons as compared to those of the mildly affected tendon and the healthy tendon at the same displacement. For example, at 0.1 mm, forces are ≥5 N for the case of severe trigger finger, whereas, for the mildly affected and healthy tendon, the forces are 3.6 and 1.83 N, respectively. Healthy and affected tendons followed a linear stress–strain relationship until triggering, with a Young's modulus of approximately 1.5 MPa [24].

For the EDC tendon (Figure 4), the increase in force in the mildly affected tendon is less as compared to the equivalent condition for the FDS tendon (Figure 5). For example, a large increase in the measured force in the mildly affected tendon only noticeably increases at 0.38 mm for the EDC tendon, whereas, for the FDS tendon, the increase is immediate.

#### *2.3. Position B—Abduction*

Displacement of the tendons themselves was negligible during abduction for position B. For this motion, measuring the displacement of each node from its initial position provided much clearer results. Comparing the force–displacement graph for position B (Figure 6) to that of the extended finger in position A (Figure 3), the finger is able to undergo a larger displacement before an increase in force is observed. For the severe case of trigger finger in position B, force starts to increase rapidly at a displacement of 6.98 mm and at 7.53 mm for the mild trigger finger. Additionally, the rate of increase in force for severe trigger finger does not differ from that of the healthy tendon until 0.625 s, whereas the equivalent point was observed at 0.3125 s in position A.

#### *2.4. Position C—Hyper-Extension*

For position C, up to the point at which locking occurs in the severely affected tendon at 0.08 mm (Figure 7), measured force increases rapidly, similar to the force increase in position A (Figure 3). The change in force gradient was noticeably lower for the mildly affected tendon when approaching the 5 N limit, with an initial rise at 0.18 mm.

A "toe" region is observed in the stress–strain curves (Figure 8) of the hyper-extension movement of the finger before the relationship becomes linear. The mildly affected tendon can withstand a greater amount of strain over a longer period before triggering occurs. The mildly affected tendon reached a strain of 0.26% at the maximum stress of 0.78 MPa, whereas the severely affected tendon only reaches a strain of 0.059%.

#### *2.5. Cross-Sectional Area and Tension*

The cross-sectional area (CSA) of the tendons decreased when loads were applied (Table 1). As a result of the severely affected tendon's force increasing more rapidly before movement is restricted, the CSA is much greater than that of the mildly affected tendon, which has a much greater capacity to extend before finger movement is restrained. This is also the case for the tension in the tendons presented in Table 2. Tension in the mildly affected tendon in both positions was more than double that in the severely affected tendons, due to their greater capacity to extend.


**Table 1.** The calculated values for cross-sectional area, comparing positions A and C. Using Equations (1)–(7).

**Table 2.** The calculated values for tension, comparing positions A and C. Using Equations (1)–(7).


#### **3. Discussion**

This study has highlighted a clear increase in exerted force on tendons restricted by trigger finger when compared to healthy tendons under the same range of motion. The force analysis further indicates that the more severe the condition, the greater the stress induced in the tendon.

As expected, tension was higher when the tendons were under hyper-extension as compared to extension (position C vs. position A). Even for the healthy tendon, tension for position C (4.70 N) was measured to be more than double that at position A (2.00 N). This is in alignment with evidence found in the literature that demonstrates that hyper-extension injuries can occur when tendons or ligaments become overstretched [11]. In the case of trigger finger, greater tension was observed in the FDS tendon for position C as compared to position A (1.13 N greater under mild trigger finger). Further tendon injury could occur during hyper-extension as a result of trigger finger; tendons could become ruptured or separated from the bone, resulting in disrupted muscle function and joint instability.

For validation of the healthy tendon, several studies have been reviewed. Yang et al. [25] and Tanaka et al. [26] both tested this using tendons taken from fresh cadavers, whereas Kursa et al. [27] and Edsfeldt et al. [28] carried out testing during open carpal tunnel surgery. These studies all reported on loading forces during flexion. Loading forces on each joint in full flexion in the study by Yang et al. ranged from 1.69 to 7.93 N [25]. The forces in the study by Kursa et al. [27] ranged from 1.3 to 4 N in FDP tendons and 1.3–8.5 N in FDS tendons, and Edsfeldt et al. [28] reported forces of up to 13 N. Although the lock constraints may have meant that the model in our study has under-predicted the highest values for force, the results for the healthy FDP observed a trend which appears consistent with the maximum force values reported in the literature.

There are limitations to using cadaver models; for instance, specimens might be embalmed or treated with chemicals to prevent degradation, embalming may increase stiffness of the tissue [29,30], and treatments such as dehydration [31] and cross-linking (e.g., using glutaraldehyde) alter the physical and mechanical properties of tissues. If samples undergo freeze-thaw cycles, this too may alter their mechanical properties [32,33]. In the case of cross-linking, chemicals such as glutaraldehyde reduce degradation through the process of cross-linking of collagen, with a side-effect of increased stiffness, though this may depend on the state of crimp of the collagen which has undergone cross-linking [34].

When considering the forces acting on the FDS tendon during trigger finger, the results in this study were compared with Lu et al. [23]. The results from this study are particularly meaningful as it is the only study detailing the direct correlation between trigger finger and its effect on the forces acting on the tendon. External force increased gradually as extension angle increased; as triggering occurred, there was an abrupt increase in force, with the maximum force reaching 5.4 N. This coincides with results from this study, with peak values of around 5 N, which implies that any limitations in using connector elements may account for less than 10% of peak predictions. Therefore, this model is in agreement with the results found in the literature for extension.

One of the most common treatments for trigger finger is splinting; the affected finger is tied to a splint to restrict movement in flexion and extension while not restricting movement in abduction and adduction. For both the mildly and severely affected tendons, a displacement of 6.9 mm was reached before a sudden increase in observed force. These findings imply that there may be value in early treatment and the necessity of healthy tendons, providing further understanding of the impact of trigger finger to avoid the need for surgery. It is noteworthy that the current conservative treatment is splinting and physiotherapy; therefore, there is clear scope for innovation in this field. Active devices could be developed which enable the appropriate loading of tendons within the hand, potentially with a limit on loading/extension as necessary, e.g., to prevent hyper-extension, etc. One option for an active, external prosthetic device could be the use of electroactive polymers. There is also the possibility to combine this technology with micro-electro-mechanical systems to sense loading. Such a device could potentially be used at home, with data recorded and logged so that, during visits to the clinic, data could be evaluated—for instance, using a radio frequency identification (RFID) tag.

There is agreement in the literature that tension in the FDP tendon is greater than that in the FDS tendon for healthy fingers [25,27]. More specifically, in a study by Lu et al., tension generated during passive extension modelling was estimated in the FDP tendon as 1.41 to 22.93 N compared to 0.78 to 11.97 N in the FDS tendon [23]. If the FDS tendon is inflamed, as experimented with in this study, the overall strain on the FDP tendon would be greater, leading to significant long-term damage. Larger moments are necessary about the joints with trigger finger, so if there is an increase in repetitive high tendon loads, further deformation may occur. Tendons may also suffer elongation with sustained loads. There is less strain on both flexor tendons with mild trigger finger.

Ultimately, the intention of this paper is partly to encourage innovation for prosthetic devices to treat trigger finger, by providing a framework for initial stage development in silico. Patient specific models [35] can be useful to tailor any technologies to individuals. Alternatively, there is scope to scale the model used in this study to more quickly enable the clinical assessment of tension for a given individual; such scalable models have been of value in other areas of orthopaedics [36]. Setting up these types of models is feasible by producing scripts which generate input files directly and request input data such as boundary conditions in a specific format (e.g., .dat files) to then enable the FEA software to perform the numerical solution. Boundary conditions for models can also use boundary conditions specific to an individual [37]. One advantage of the generation of any computer-aided design model is that it can be 3D printed [38], which can also be useful for evaluation or interaction with patients when explaining the condition.

#### **4. Materials and Methods**

#### *4.1. Geometry*

A geometric model of a human skeleton was sourced [39] (Figure 9) and imported into computer-aided design software (SolidWorks, Dassault Systémes, Vélizy-Villacoublay, France), from which the hand bone structure was extracted. The bones were scaled using Solidworks; scaling was implemented so as to match the dimensions of an adult female available from the literature [40] (Table 3). The distal, middle and proximal phalanges and the metacarpal bone of the index finger were saved as separated parts before being imported into ABAQUS (Dassault Systems, Providence, RI, USA) as three-dimensional (3D) deformable components. The tendons and ligaments were then modelled using connector elements and constraints.

**Figure 9.** CAD model of the right hand used for simulations.

**Table 3.** The phalangeal and metacarpal lengths used in the CAD model.


#### *4.2. Material Properties*

The material properties for cortical bone [41] were assigned to each component, as detailed in Table 4. Ligaments and tendons are comprised of bundles of closely packed collagen fibrils [42–45], organised in parallel to resist strong tensile loads. Therefore, tendons, for instance, display hyperelastic properties, and as the finger is straightened in response to an applied load, the tendons start to deform in a linear fashion and become aligned [45].

**Table 4.** The properties of cortical bone implemented into the ABAQUS model; from the literature [41].


Tendons were assumed to be incompressible, with no change in volume. Additionally, tendons were modelled as undergoing frictionless motion through the tendon sheath. However, where trigger finger was included in a simulation, the motion was restricted, as explained in Section 4.3.2. The path of the tendon was assumed to follow a straight line along the surface of the bone between two points in the tendon network. The loads experienced by a tendon were assumed to be distributed uniformly throughout the tendon network. Tendons were modelled using material properties from the literature [46]; this data was inputted directly into the ABAQUS, such as the data shown in Figure 10.

#### *4.3. Model Set-Up*

#### 4.3.1. Joint Orientation

A kinematic model of the hand can be mathematically approximated as a number of revolute joints that are linked together. The index finger model is based on methods commonly used in the literature [1,47,48]. The distal (DIP) and proximal interphalangeal (PIP) joints have one degree-of-freedom (DOF) and are modelled as frictionless hinge joints capable of flexion-extension motion. The metacarpophalangeal (MCP) joint represents two DOF and is modelled as a frictionless saddle joint capable of flexion-extension and adduction-abduction. Coordinate systems were defined for each bone in the index finger with respect to a common inertial frame of reference to provide orientation of the joints and tendon configuration. A coordinate system for the distal, middle and proximal phalanx was used [49]. Each system is located in the centre of rotation in the convex articular surfaces of the phalangeal and metacarpal heads. The x-axis is projected along the shaft of the bones, with the y-axis projected dorsally and the z-axis projected radially for the right hand (note: these frames of reference are local to each individual bone).

**Figure 10.** Sample force-extension data used to simulate tendons. (**a**) Full data set used from the literature [46]; (**b**) portion of the data relevant to the loading range within the simulations solved.

Initially, the phalanges were arranged in a flexed position [48] to resemble how the finger may be immobilised in the "trigger" position (Figure 11). Constraints were applied to ensure that each bone moved in the appropriate DOF and rotated accordingly within each coordinate system. A reference point was assigned at the centre of each convex surface of the bones. A coupling constraint was used which provides a coupling between a reference point and a group of nodes, with the necessary DOF applied between the articular surfaces. This ensures that the distal bone to the convex surface would only rotate about this surface. A general multi-point constraint (MPC) was used between each bone to ensure movement between bones was coordinated. The MPC was applied between the reference point and articular surface of each of the bones. This meant that only two displacement/rotation boundary conditions were necessary for the whole model, one to extend the finger and another to simultaneously straighten the distal phalanx.

**Figure 11.** The initial flexed position of the index finger with the locations of multi-point constraints (MPCs) labelled.

#### 4.3.2. Tendon and Ligament Modelling

Tendons and ligaments were simulated within the model as connector elements and constraints. Following the meshing of the bone structures, nodes were placed on the mesh at points of attachment in the tendons. To replicate the tendons being constricted by ligaments along the bone, connector elements were joined between the nodes on the bone. Nodes were spread evenly along the palmar and posterior aspects of the finger (Figure 12a).

**Table 5.** A table presenting the different movement-types investigated.

(**b**)

**Figure 12.** Connector elements used. (**a**) The positioning of the nodes and connector elements on the bones (the right image shows the nodes for the FDS tendon). (**b**) The locations of the lock constraints on the connector elements on the FDS and EDC tendons. Please refer to Table 5 for the movements mimicked.

Connector elements were placed so as to mimic the FDP tendon extending from the distal phalanx to the proximal end of the metacarpal, with the FDS tendon attaching to the middle phalanx in a forked arrangement. The FDS tendon is also constrained by the same set of nodes as the FDP tendon to the proximal metacarpal, both on the palm side. The FDP, FDS and EDC tendons were all modelled in the simulation, with the EDC tendon extending along the full length of the back of the finger, similarly held in place by nodes. A lock constraint as part of the connector element was applied on the FDS and

EDC tendons between the two nodes on either side of the MCP joint (Figure 12b). This is a common area in which tendons can experience irritation and consequently inflammation and restriction [50].

Lock constraints restrict movement after a set displacement; the displacement is the change in distance from one node to another. In this study, an estimate was made for the displacement of the lock based on the length of a flexor tendon in an index finger of an adult female. Displacement was set at 0.1 mm for severe trigger finger for relatively no movement and 0.5 mm for mild trigger finger to allow for some movement; 0.1 mm was used to avoid modelling artefacts which became evident when 0 mm was used in preliminary models, and 0.5 mm was used to determine how greater movement would alter tension and the timing of tension which would develop in tendons. To model a lock constraint in ABAQUS, two parameters are needed: firstly, a parameter being constrained (in this case, displacement at 0.1 or 0.5 mm) and a parameter which enforces this constraint (e.g., force). Preliminary models found that a 5 N load suitably enforced the lock constraint on displacement (of both 0.1 and 0.5 mm).

The movements evaluated are outlined in Table 5. For each movement type, results were taken for both severe and mild cases of trigger finger. Initially, the FDP and EDC tendons were modelled without any of the restrictions of trigger finger, whereas the FDS tendon became locked after the specified displacement. Starting from the initial flexed position, the finger was extended for position A, abducted for position B and hyperextended for position C.

#### 4.3.3. Boundary Conditions

The movements outlined in Section 4.3.2 were feasible through the use of the boundary conditions outlined in Table 6. The metacarpal bone is stationary throughout the running of the simulation; therefore, an initial boundary condition of encastre was applied. In all positions, the finger started in the initial flexed position. For position A, the middle phalanx was rotated 96◦ about the z-axis until the finger was aligned to a straight orientation. The distal phalanx followed this direction of motion due to the MPC constraints and was simultaneously rotated 61◦ about the z-axis. The proximal phalanx was rotated 40◦ about the y-axis, simulating abduction for position B. After returning to the initial position, the finger was hyper-extended, rotating the proximal phalanx 45◦ about the z-axis.


**Table 6.** A list of the boundary conditions used in ABAQUS. U1, U2 and U3 specify movements along Table 1. UR2 and UR3 specify rotation about the *x*, *y* and *z* axes, respectively.

#### *4.4. Analysis*

The cross-sectional area of a flexor tendon mid-section can range from 8.36 to 14.44 mm<sup>2</sup> [51,52] for an index finger. For this simplified model, the assumption made was that the FDP, FDS and EDC all have an initial average circular cross-sectional area, A, taken from the literature [53] (Table 7).

**Table 7.** The initial values used for initial cross-sectional area, initial length and Young's modulus used in Equations (1)–(7).


Assuming homogenous material properties, uniform stiffness k for collagen can also be adopted along the tendon length [53]. For convenience, the mean slack length of the tendon will equal the mean initial tendon length *L*0. Equations (1)–(3) have been adapted from Freij et al. [54], Equation (4) from Vigouroux et al. [22] and Equations (5)–(7) from Young et al. [53] to calculate the actual cross-sectional area a, true stress σ and stiffness after each position. Where λ is the stretch ratio, *L* is the actual length and *F* is force given in ABAQUS.

$$
\lambda = \frac{L}{L\_0} \,\tag{1}
$$

$$a = \frac{A}{\lambda'} \tag{2}$$

$$
\sigma = \frac{F}{a'} \tag{3}
$$

The tension, *T*, in the tendons can be estimated by relating force on the tendon to its elongation. The tension, *T*, in the tendons can be estimated by relating force on the tendon to its elongation through displacement using a quadratic function (Equation (4)).

$$T = k(L - L\_0)^2,\tag{4}$$

Stiffness *k* is estimated using Equation (5), where *E* is the Young's modulus of the tendon, which has previously been reported as being approximately 1.5 MPa [24], where ε is strain and Δ*L* is extension.

$$k = \frac{AE}{L\_0},\tag{5}$$

$$
\varepsilon = \frac{\Delta L}{L\_0} \,\prime \tag{6}
$$

$$E = \frac{\sigma}{\varepsilon'} \tag{7}$$

Equations (1)–(7) were used to calculate tension in the tendon, stress and strain. The displacement and force of the connector element in the model, between the two nodes either side of the MCP joint, have been used when evaluating the equations during movements A, B and C (Table 6) for the healthy FDP tendon and the mildly and severely affected FDS tendon. Additionally, for movement A, analysis included the EDC healthy, mildly and severely affected tendons. Table 7 contains the initial values used for Equations (1)–(7).

#### *4.5. Mesh and Solution*

The analysis was performed using ABAQUS. A dynamic, explicit solution type was used with one to two steps. A time period of one second was applied for each step with boundary conditions using a tabular amplitude. The geometric complexity of the bones favoured the use of an automatic tetrahedral mesh, deemed acceptable as the bones were not expected to undergo large deformation. Mesh convergence consisted of increasing the node seed size in equal steps and identifying the size at which it did not alter predictions of tension in tendons. Mesh convergence occurred at a seed-size of 4.5, which resulted in a mesh with 5123 elements (solution time: ~14.2 min). It should be noted that only the bone structures are actually meshed and, given the loading involved, these undergo deformation which is negligible (in essence, acting as rigid bodies); this is why mesh convergence has occurred using a low number of elements.

#### **5. Conclusions**

This study demonstrates the effects that trigger finger has on the extensor mechanism, and it predicts tendon loads as caused by trigger finger as compared to a "healthy" control case. There appears to be a substantial increase in tension during hyper-extension during trigger finger. It is hoped that the model presented could be used as a framework to enable more advanced treatment methods to be developed than currently available (e.g., prosthetic devices for rehabilitation).

**Author Contributions:** Conceptualization, D.M.E. and C.G.B.; methodology, H.I.R. and D.M.E.; software, H.I.R.; validation, H.I.R.; formal analysis, H.I.R.; investigation, H.I.R. and D.M.E.; data curation, H.I.R.; writing—original draft preparation, H.I.R. and D.M.E.; writing—review and editing, D.M.E., H.I.R. and C.G.B.; supervision, D.M.E. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Editorial* **Worldwide 3D Printers against the New Coronavirus**

**Luca Fiorillo 1,\* and Teresa Leanza <sup>2</sup>**


Received: 30 May 2020; Accepted: 4 June 2020; Published: 5 June 2020

**Abstract:** The pandemic caused by the new coronavirus has placed national health systems of different countries in difficulty, and has demonstrated the need for many types of personal protective equipment (PPE). Thanks to the advent of new three-dimensional printing technologies, it was possible to share print files (using stereolithography (stl)) quickly and easily, improve them cooperatively, and allow anyone who possessed the materials, a suitable 3D printer and these files, to print. The possibility of being able to print three-dimensional supports, or complete personal protective equipment has been of incredible help in the management of COVID-19 (Coronavirus Disease 2019). The times and the relatively low costs have allowed a wide diffusion of these devices, especially for the structures that needed them, mainly healthcare facilities. 3D printing, now includes different fields of application, and represents, thanks to the evolution of methods and printers, an important step towards the "digital world".

**Keywords:** coronavirus; COVID-19; 3D-printing; DPI; protection; public health

Three-dimensional printers and millers are now widely used in the world; both in the medical field and in other areas such as electronics, engineering, construction, and military fields [1]. Technological development has made it possible to print (or mill) different materials, with multiple characteristics, and with good resolutions and reliability compared to analogue techniques [2–4]. 3D (3 dimensional) printing (or additive manufacturing) allows the creation, starting from a digital model, of three-dimensional physical objects, by depositing, layer by layer, of overlying materials. 3D printing dates back to the 80s. Beginnings could be considered with stereolithography (rapid prototyping and STL (STereo Lithography) format), followed by sintering (selective laser sintering), to arrive at fused deposition modeling or 3D printing with molten material [5,6]; it was thanks to the three dimensional printing that it became possible to print in color, while with the Electron beam melting, or even electron beam fusion, it was possible to obtain metal objects with a high density [7,8].

At the end of 2019, scientists isolated a new coronavirus in these subjects, designated SARS-CoV-2 (Severe Acute Respiratory Syndrome—Coronavirus-2), found to be similar to at least 70% of its gene sequence to that of SARS-CoV. Patients experience flu-like symptoms such as fever, dry cough, tiredness, difficulty breathing [9–15]. Certainly, the most common methods of diffusion of the virus involve the spread of infected droplets at a distance, through coughing, sneezing, or simply speaking. In more severe cases, often found in subjects already burdened by previous pathologies, pneumonia develops, acute renal failure, up to even death. 3D printing has found wide application in the medical field also in the Covid-19 (Coronavirus Disease—2019) emergency period. In fact, given the difficulty in finding official health supplies, thanks to this technology, many people have mobilized in a completely autonomous way to find concrete solutions. The whole world of "markers" and digital companies has exploited 3D printers in order to make up for these shortcomings by creating spare parts, fittings and

compatible tubes for medical instruments in record time, useful for dealing with emergencies [16–19]. Recently, doctors, running out of valves for respiratory intensive care equipment and unable to purchase them from companies, needed to find a solution to save the lives of hospitalized patients.

This impactful event shone the spotlight on the activity within the community made up of makers and producers, which for some time had started to make a contribution to the health emergency, starting to organize itself to produce materials missing or that it would have been better to manufacture on site to avoid delivery delays. In fact, to print in 3D it is not enough to connect a machine to the internet. However, you also need:


In a short time, makers and big producers came to organize themselves to have everything quickly. CAD (Computer Aided Design) files of all types have sprung up at the makers' sites, from valves to face masks. On Facebook® the communities and groups in which people recommend the best materials, the most effective design and so on have multiplied too. University researchers are contributing to efforts in various centers of excellence. Finally, companies that supply pieces to complement those printed, such as fabrics or screens for protective masks, have accelerated their production and donated materials to the Italian regions (as Decathlon® did with its snorkeling masks). In short, the market has been populated with alternatives and solutions, giving life to an extraordinary offer, albeit very varied in terms of quality [20–26].

Especially when it comes to medical devices and personal protective equipment, European consumer protection legislation is very strict and requires a long process of certification before being placed on the market. For example, some protective equipment may require approval by the Food and Drug Administration (FDA). Therefore, it must be ensured that companies that deal with additive manufacturing have a pass to continue their activity even during the lockdown period. At the international level, the flow of data (and therefore the design of 3D printed objects) should continue to be free of localization policies, barriers and duties. 3D printing has undoubted advantages when it comes to small productions, custom objects or complex designs. The additive manufacturing cancels the adaptation and configuration times of the machines (such as those for the creation of new molds), and allows the creation of complex and customized pieces without splitting up costs and production times [27]. It therefore remains to be understood how, once the emergency has passed, people may be able to take advantage of the specific benefits of this technology and whether its use in crisis situations will push for a more widespread adoption.

**Author Contributions:** Validation, T.L., project administration, L.F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** In this COVID-19 emergency period, thanks go to all clinicians and researchers who everyday risk their lives for research.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


#### *Prosthesis* **2020**, *2*


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **3D Printing beyond Dentistry during COVID 19 Epidemic: A Technical Note for Producing Connectors to Breathing Devices**

#### **Leonardo Cavallo 1, Antonia Marcianò 2, Marco Cicciù 3,\* and Giacomo Oteri <sup>3</sup>**


Received: 30 March 2020; Accepted: 3 April 2020; Published: 7 April 2020

**Abstract: (1)** Background: To mitigate the shortage of respiratory devices during the Covid-19 epidemic, dental professional volunteers can contribute to create printed plastic valves, adapting the dental digital workflow and converting snorkeling masks in emergency CPAP (continuous positive airways pressure) devices. The objective of this report was to provide the specific settings to optimize printing with the 3D printers of the dental industry. **(2)** Methods: In order to provide comprehensive technical notes to volunteer dental professionals interested in printing Charlotte and Dave connectors to breathing devices, the entire digital workflow is reported. **(3)** Results: The present paper introduces an alternative use of the dental Computer Aided Design/Computer Aided Manufacturing (CAD/CAM) machinery, and reports on the fabrication of a 3D printed connection prototypes suitable for connection to face masks, thereby demonstrating the feasibility of this application. **(4)** Conclusions: This call for action was addressed to dentists and dental laboratories who are willing to making available their experience, facilities and machinery for the benefit of patients, even way beyond dentistry.

**Keywords:** Covid-19; CAD/CAM; 3D printing; dental prostheses; resin printed device

#### **1. Introduction**

The pandemic outbreak of a severe acute respiratory syndrome (SARS) associated with the novel coronavirus (2019-nCoV) poses a serious public health risk due to the high number of patients demand for ICU admission and mechanical ventilation [1,2].

To date (28 March 2020) in Italy 26,676 patients are hospitalized without mechanical ventilation, and 39,533 are advised to recover at home without continuous medical care, although in many of these cases, a support for spontaneous ventilation is needed [3].

CPAP (continuous positive airways pressure) allows the insufflation of air and oxygen at positive pressure in a continuous and non-invasive way for the duration of the respiratory cycle. The CPAP can be delivered through a mask (facial or nasal), flow and oxygen dispenser or with a mechanical fan. The choice of device depends on the patient's clinical condition, the environment in which it is delivered and the technological resources available [4].

In emergency settings, CPAP is an important alternative to invasive mechanical ventilation. Nevertheless, it can be used in several acute and chronic respiratory diseases, and also at home with a portable oxygen source, as provided in obstructive sleep apnea treatment [5].

At the moment, given the serious coronavirus pandemic, the majority of Italian hospitals do not have sufficient equipment to assist patients affected by respiratory failure, by reducing alveolar compression and supporting breathing [6].

In order to find an urgent solution to the need of respiratory devices, an Italian engineer (Dr. Eng. C. Fracassi) has ideated a respiratory device consisting of a commercially available snorkeling face mask *Easybreath* (Decathlon-Villeneuve-d'Ascq, France) in which the respiratory tube is replaced with a plastic support suitable to be connected to medical oxygen supply pipes.

The project of the connector has been designed by the Italian company ISINNOVA (Brescia, Italy), which has released on the web the related "standard triangulation language" STL files for free.

Consequently, in order to have the facial mask available for the conversion into a CPAP, it is necessary to apply the specifically designed fitting component consisting of two connection pieces that are printable with modern three-dimensional printers (3D).

The developed respiratory device is the result of the application of two 3D printed plastic valves *"Charlotte*" and *"Dave*" to the *Easybreath mask* (Model *Subea 1*).

The entire system set-up is made up as follows:


Considering the growing Covid-19 epidemic and the consequent exceptional case of necessity, the Italian Ministry of Health allows the use of these non-certified biomedical devices for compassionate care.

The application of an information procedure and a specific patient's informed consent is requested before using these devices [8].

Because of Covid-19, hospitals are urgently requesting breathing devices; groups of volunteers working in research centers, companies, individuals and among them also dentists and dental technicians have joined together to quickly create 3D printing fittings.

The present paper introduces an alternative use of the dental CAD/CAM machinery, reporting on the fabrication of a 3D printed connections prototype suitable for connection to snorkeling face masks, demonstrating the feasibility of the application. 3D printing companies act as central hubs connecting makers and hospitals in need by crowdsourcing a list of professional additive manufacturing (AM) providers who have suiTable 3D printers. Dentists and dental laboratories who are willing to making available their experience, facilities and machinery for the fight against the coronavirus can sign up at https://www.3dsystems.com/covid-19-response#signUp [9].

The objective of this report was to present the specific workflow to be applied for printing the connectors with Dental 3D Printers that meet the reported setting requirements. Medical-grade materials must be used.

#### **2. Results**

In order to provide comprehensive technical notes to volunteer dental professionals interested in printing Charlotte and Dave connectors to breathing devices, the entire digital workflow is reported.

#### *Step by Step Procedure*

The steps leading to the resulting pieces are described below.

*Prosthesis* **2020**, *2*

The STL file available by ISINNOVA srl. of the "*Charlotte*" valve and "*Dave*" valve is imported to the 3D printer software(PreForm 3.4.3 Formlabs Inc.) (Figures 1 and 2).

**Figure 1.** STL file of "*Charlotte*" valve (made available by ISINNOVA srl.).

**Figure 2.** STL file of "*Dave*" valve (made available by ISINNOVA srl.).

The corrected position of the pieces in the printer plate (Figure 3) and the number, position, height and diameter of the supporting pins are checked according to instructions.

**Figure 3.** "*Charlotte*" and "*Dave*" valves leaning on the printer terminal plate.

The print is then launched at a predetermined operating time (5 h 15 min).

At the end of the printing, (Figure 4) a post-production phase is required.

The polymerized components are removed from the printer and washed in **Isopropyl alcohol** in an ultrasound tank for 5–15 min.

Subsequently, they are cured and dried in the UV curing system Meccatronicore BB Cure Dental (Meccatronicore, Trento, Italy).

**Figure 4.** Printed vales external and internal vision.

All supporting pins are removed, and the external surfaces of the plastic devices are finished using conventional dental methods, with rotating burs and brushes (Figure 5).

**Figure 5.** Finishing.

The last step is cleaning with a broad-spectrum disinfectant hydroalcoholic solution (**Bactisan Spray, Amedics**).

Finally, the valves, stored in sterilization tubing to avoid contamination, are ready for delivery, since correct adaptation to the mask is ensured. (Figures 6–8).

**Figure 6.** Insertion of "*Charlotte*" valve to the mask.

**Figure 7.** Printed connector detail comparison to *Subea 1.*

**Figure 8.** Comparison between "*Charlotte*" and *Subea* 1 insertion.

#### **3. Discussion**

The severe and widespread Covid-19 pandemic puts many people's lives at risk all over the world. The insufficient number of beds in Intensive Care Units associated with the huge demand for assisted breathing devices cannot be met by factory supplies in a short time.

Alternative and emergency respiratory apparatuses can help the breathing of many patients who are affected by coronavirus.

However, it is noteworthy to mention that in Covid-19 patients with acute respiratory failure, CPAP may not be an adequate treatment. Therefore, since it is difficult to predict what these cases are, the decision to try this treatment is up to the intensive care specialist and they need to provide close monitoring, including preparations for prompt intubation.

3D printing translates computer-aided design (CAD) virtual 3D models into physical objects. 3D printing is used in the manufacturing industry, medical and pharmaceutical research, drug production, clinical medicine and dentistry, with implications for precision and personalized medicine [10,11].

The term 3D printing with the alias Customized Additive Manufacturing (AM) is used to describe the same general manufacturing principle that builds objects layer by layer.

AM techniques include vat photopolymerization (stereolithography), powder bed fusion (SLS), material and binder jetting (inkjet and aerosol 3D printing), sheet lamination (LOM), extrusion (FDM, 3D dispensing, 3D fiber deposition and 3D plotting) and 3D bio printing [12].

With the advent of computer-aided design/computer-aided manufacturing (CAD/CAM) protocols, it became quite popular in dentistry, especially for implant prosthodontics [13,14].

Dental professionals have a deep awareness of digital workflow for 3D printing, since the use of it to build dental models, fixed prostheses, full-arch implant supported rehabilitation and others is nowadays routine in the daily dental practice.

Volunteer dental professionals can contribute to creating printed plastic valves, adapting the dental digital workflow and converting snorkeling masks in emergency CPAP devices.

The role of the dentist and the dental laboratory is only limited to making available their experience, facilities and machinery for helping doctors and patients, even way beyond dentistry [15–17].

#### **4. Materials and Methods**

The free STL files of Charlotte and Dave valves were downloaded from the link: https://drive. google.com/drive/folders/14Q3TEl5JVeN2QpDpKo1AIx\_wnGeolKlK?usp=sharing.

The low force stereolithography (LFS) Formlabs Form 2 printer (Formlabs Inc., Somerville, MA, USA) was used. Stereolithography is an additive manufacturing process that, in its most common form, works by focusing an ultraviolet (UV) laser on to a vat of photopolymer resin; the resin is photochemically solidified and forms a single layer of the desired 3D object from the computer-aided design (CAM/CAD) software. This process is repeated for each layer of the design until the 3D object is complete [13].

In the design phase, the option "automatically generates everything" allows us to take advantage of the pre-defined settings for the creation of the supports.

The following basic settings were used:

Density: 1.00 Size of the contact points: 0.90 Internal supports: on Spacing from the plane: 5.00 Inclination multiplier: 1.00 Height above the base: 5.00 Base thickness: 2.00 Layer thickness: 0.1 mm Print time: 5 h 15 min Layers: 932 Volume: 60.52 mL

The printer is used in "open mod" activity to allow the use of resins other than the original ones supplied by Formlabs.

In this case, the employed printing material was NextDentTM C&B (NextDent B.V., 3769 AV Soesterberg, Netherlands), a micro filled hybrid (MFH) class II a printing material suitable for medical devices, biocompatible and CE certified in accordance with Medical Device Directive 93/42/EEC, listed by the FDA and registered in various other countries.

For this prototype, the color A 3,5 was chosen.

This material reflects the characteristics of the ideal material for this use, which should be odorless, biocompatible, biocomposable and relatively flexible to easily connect with the mask component.

**Author Contributions:** L.C. and A.M. equally conceived and designed the article; L.C. carried out the devices. G.O. drafted the manuscript and coordinated the study; M.C. critically revised the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** Declare none.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

1. COVID-19. Available online: https://www.ecdc.europa.eu/en/novel-coronavirus-china (accessed on 1 March 2020).


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Prosthesis* Editorial Office E-mail: prosthesis@mdpi.com www.mdpi.com/journal/prosthesis

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18