Introduction
Traumatic tympanic membrane perforation is common. It tends to heal spontaneously, and the traditional therapy involves keeping the ear dry.Reference Griffin 1 – Reference Gladstone, Jackler and Varav 5 However, the time to spontaneous closure is long and the closure rate of large perforations is low. Many treatments are used to shorten the closure time and improve the closure rate, including patching, edge approximation, topical growth factors and myringoplasty. Ear drops can transfer bacteria from the external auditory canal into the middle ear, causing middle-ear infection, which impairs healing. The best therapy for traumatic tympanic membrane perforation is not clear.
In recent years, moist therapy has gradually been accepted. Clinical and experimental studies have shown that growth factors (including epidermal growth factor and basic fibroblast growth factor) and various blood preparations facilitate tympanic membrane regeneration. The effects of growth factors are not included in this review; rather, it focuses on the role played by a moist environment created by the application of solutions not containing growth factors in tympanic membrane perforation repair. In addition, some features of moist therapy are discussed.
Materials and methods
We used PubMed to search the literature on August 2016, and identified all of the published studies on tympanic membrane perforations.
First, we retrieved 147 articles using the following key term combinations: ‘solution’ and ‘tympanic membrane’; ‘eardrops’ and ‘tympanic membrane’; ‘ofloxacin otic drops’ and ‘tympanic membrane’; ‘ointment’ and ‘tympanic membrane’; ‘topical application’ and ‘tympanic membrane’; ‘EGF’ and ‘tympanic membrane’; or ‘bFGF’ and ‘tympanic membrane’. The articles were screened by reviewing the titles or abstracts based on our inclusion and exclusion criteria. The inclusion criteria were: English-language articles, treatment of patients or animals with traumatic tympanic membrane perforations, at least one treatment (control or otherwise) featuring a solution that did not contain growth factors, and a between-group comparison of healing outcomes (healing rate and/or healing time). The exclusion criteria were: review or commentary articles, treatment of chronic tympanic membrane perforation, and performance of a histological study (only) on the effects of growth factors (or blood products).
Second, a total of 1330 articles were identified by an initial search using the word ‘myringoplasty’. These papers were screened by reviewing the titles or abstracts in the context of the inclusion and exclusion criteria. The inclusion criteria were: a human clinical study, the use of type I tympanoplasty (myringoplasty), a comparison of success rates in terms of the effects of wet and dry middle-ear mucosae and discharge, and publication in the English language. The exclusion criteria were: treatment of traumatic tympanic membrane perforations, an experimental study, and comparison of the success rates of different graft materials.
Results
Fourteen articles used solutions lacking growth factors to treat acute tympanic membrane perforations, and were included in this review. Table I summarises the articles that compared the outcomes of growth factors and placebos on traumatic tympanic membrane perforations.Reference Fina, Baird and Ryan 6 – Reference Vrabec, Schwaber, Davidson and Clymer 12 Of the 14 studies, 5 used phosphate-buffered saline, glycerol, or sodium chloride 9.11 mg/ml and benzalkonium chloride 0.0479 mg/ml as control solutions when growth factor containing solutions were used to treat traumatic tympanic membrane perforations.Reference Fina, Baird and Ryan 6 , Reference Chauvin, Bratton and Parkins 7 , Reference Fina, Bresnick, Baird and Ryan 10 – Reference Vrabec, Schwaber, Davidson and Clymer 12 One study used an insulin solution to repair tympanic membrane perforations.Reference Araujo, Murashima, Alves, Jamur and Hyppolito 13 Another study used saline as a control when embryonic stem cells were used to treat traumatic tympanic membrane perforations.Reference von Unge, Dirckx and Olivius 9
Table I Summary of articles comparing outcomes of growth factors and placebos on traumatic tympanic membrane perforations

FGF = fibroblast growth factor; EGF = epidermal growth factor; PBS = phosphate-buffered saline
Table II summarises the articles that compared the outcomes of non-growth factor solution on traumatic tympanic membrane perforations.Reference Araujo, Murashima, Alves, Jamur and Hyppolito 13 – Reference Akkoc, Celik, Arslan, Demirci, Hucumenoglu and Caydere 20 Five studies used Ofloxacin Otic Solution to repair traumatic tympanic membrane perforations.Reference Lou, Lou, Tang and Xiao 14 – Reference Buyten, Kaufman and Ryan 18 One study used an ointment to repair traumatic tympanic membrane perforations.Reference Amadasun 19
Table II Summary of articles comparing outcomes of non-growth factor solutions on traumatic tympanic membrane perforations

EGF = epidermal growth factor; FGF = fibroblast growth factor
Nine articles compared the myringoplastic outcomes of dry and moist middle-ear mucosae, and were included in this review.Reference Onal, Uguz, Kazikdas, Gursoy and Gokce 21 – Reference Vijayendra, Rangam and Sangeeta 29 Of these, only one study suggested that a moist middle-ear mucosa increased the failure rate of myringoplasty;Reference Onal, Uguz, Kazikdas, Gursoy and Gokce 21 the other eight studies found that a moist middle-ear mucosa did not affect the myringoplastic success rate.Reference Santosh, Prashanth and Rao 22 – Reference Vijayendra, Rangam and Sangeeta 29
Discussion
History and advantages of moisture in wound healing
It was once widely accepted that successful wound healing depended on maintaining a dry wound bed. In 1958, OdlandReference Odland 30 discovered that the speed of healing of a wound with an intact welt was faster than that of a wound with a ruptured welt. Studying the healing of skin wounds in pigs in 1962, WinterReference Winter 31 observed that if the welt was not pierced, it facilitated the movement of superficial cells and promoted rapid wound healing. The next year, Hinman and MaibachReference Hinman and Maibach 32 confirmed these findings. Additionally, ZahirReference Zahir 33 found that when the wound is kept moist, the epidermal layer grows through an exudate immediately under the blood clot. In comparison, under dry conditions, dehydration due to air exposure causes the epidermal layer to migrate through the newly formed fibrous tissue of the dermis, just underneath the wound surface, while the more superficial part of the dermis then becomes part of the scab. In 1972, Rovee et al.Reference Rovee, Kurowsky and Labun 34 proved that a moist wound heals faster than a dry wound, which was a paradigm shift toward moist wound healing that has revolutionised acute and chronic wound care.
It is now widely accepted that a moist wound environment hastens the healing of both acute and chronic wounds, and promotes the growth of new tissue.Reference Alvarez, Mertz and Eaglstein 35 – Reference Dowsett and Ayello 38 The balance of moisture is critical to wound healing. In acute wounds, a balanced moist surface facilitates the action of growth factors, cytokines and chemokines. This promotes cellular growth, collagen proliferation within a healthy acellular matrix, and the establishment of a provisional, healthy acellular wound matrix.Reference Winter 31 , Reference Okan, Woo, Ayello and Sibbald 39 – Reference Mast and Schultz 41 A moist wound environment also leads to an increased electrical gradient between the wound and wound margin and base, promoting the migration of keratinocytes into the centre of the wound base.Reference Eaglstein, Davis, Mehle and Mertz 36 This electrical gradient increases the density of platelet-derived growth factor and fibroblast growth factor receptors on fibroblasts, facilitating healing.Reference Kloth 42 In addition, in a wet wound, and in a moist wound to some extent, the increase in available nutrients and reduction in evaporative loss make the micro-metabolic balance favour repair.Reference Svensjö, Pomahac, Yao, Slama and Eriksson 43
In one study, no complications were associated with the use of a moist environment, such as wound infection.Reference Svensjö, Pomahac, Yao, Slama and Eriksson 43 In comparison, a dry wound environment promotes wound desiccation, necrosis and eschar formation, and results in poorer wound healing rates. The formation of an eschar slows the ability of regenerative cells (keratinocytes) to migrate from the wound periphery into the wound centre, hindering re-epithelialisation. In addition, necrosis enlarges the wound, necessitating additional regenerative processes. A dry environment may also delay healing by interfering with cell communication because the process relies on the diffusion of soluble factors.Reference Dowsett and Ayello 38 , Reference Svensjö, Pomahac, Yao, Slama and Eriksson 43 , Reference Vogt, Andree, Breuing, Liu, Slama and Helo 44 Overall, a moist wound environment can accelerate wound healing.
History of moist therapy in tympanic membrane regeneration
The classic treatments to achieve tympanic membrane regeneration include spontaneous healing, repeated stimulation of the edge with an acidic agent, paper patches, and myringoplasty.Reference Griffin 1 – Reference Gladstone, Jackler and Varav 5 , Reference Dunlap and Schuknecht 45 Topical ear drops were historically forbidden because they were thought to transfer the external auditory canal bacteria into the middle ear and induce middle-ear infection.
In 1947, Dunlap and SchuknechtReference Dunlap and Schuknecht 45 reported that the application of a cotton ball soaked with paraffin oil and saturated with urea ear drops incited an inflammatory response and aided closure. They removed the dried crust approximately twice a month because the crust was thought to cover the active drum margin and prevent closure mechanically. Subsequently, with the widespread use of growth factors in skin wounds,Reference Miyoshi, Kawazoe, Igawa, Tabata, Ikada and Suzuki 46 corneal injuryReference Yan, Wu, Wang, Li, Lu and Duan 47 and bone lesions,Reference Kamo, Miyakoshi, Kasukawa, Sasaki and Shimada 48 growth factors have been adopted to repair acute and chronic tympanic membrane perforation.Reference Fina, Baird and Ryan 6 , Reference Chauvin, Bratton and Parkins 7 , Reference Mondain and Ryan 49 – Reference Ishimoto, Ishibashi, Bottaro and Kaga 51
The tympanic membrane is unique in that it is suspended in air. Most reports suggest that eardrum healing depends on the proliferation and migration of epithelium from germinal centres in the handle of the malleus and tympanic annulus to close the perforation.Reference Wang, Wang and Chi 52 – Reference Makino, Amatsu, Kinishi and Mohri 54 A few studies showed that eardrum healing was similar to the healing of skin wounds in which the proliferation of granulation tissue at the perforation edges first closed the perforation, and subsequently an epithelial layer covered the granulation tissue via a fibrous layer at the perforation edge.Reference Taylor and McMinn 55 , Reference Martini, Govoni, Beghi, Fontana, Lini and Marini 56 The two different eardrum healing patterns could be dependent on the environment of the perforation edge.Reference Boedts 57
A large study of spontaneous healing of traumatic tympanic membrane perforation found that the healing of dry perforations started with the proliferation of a thin, transparent epithelium; proliferation of the fibrous layer subsequently occurred, and the new eardrum became thicker.Reference Lou, Wang and Su 58 , Reference Lou, Tang and Yang 59 The healing of moist perforations with serosanguineous otorrhoea started with the proliferation of granulation tissue at the perforation edge to close the perforation, and this was subsequently replaced by epithelium. Moist perforations had a significantly shorter closure time and better closure rate than dry perforations. In addition, the repair of a dry perforation may result in the formation of a crust and the outward migration of epithelium, thereby prolonging closure or leading to a failure to heal.Reference Lou 60 , Reference Lou 61 The proliferation of granulation tissue at the perforation edges accelerated eardrum healing. A clinical study of Gelfoam patching in traumatic tympanic membrane perforation found that the oedema and proliferation of granulation tissue at the perforation edge was more obvious and the perforation healed more quickly.Reference Lou, Tang, Chen and Xiao 62 Fagan and PatelReference Fagan and Patel 63 recommended that no attempt be made to remove the blood at the perforation edge because it makes an excellent surgical dressing.
The formation of granulation tissue is needed to accelerate eardrum healing. Nevertheless, the extent of granulation tissue is related to the wound environment, and the development of granulation tissue is more rapid in moist wounds than in dry or wet wounds.Reference Svensjö, Pomahac, Yao, Slama and Eriksson 43 , Reference Vogt, Andree, Breuing, Liu, Slama and Helo 44 In brief, a moist environment at the perforation edges not only aided regeneration of traumatic tympanic membrane perforations but also improved the success rates of myringoplasty used to treat chronic tympanic membrane perforations.Reference Park, Kim, Jin, Maeng, Yeo and Park 64 – Reference Lou, Yang, Tang and Fu 69
Moist status of middle-ear mucosa during myringoplasty
Type I myringoplasty was used to repair chronic tympanic membrane perforations caused by chronic suppurative otitis media, grommet insertion and trauma. Most of the authors considered that the operative ear had to be dry for at least three months prior to myringoplasty. A moist or wet middle ear could develop a secondary infection, resulting in failure of type I myringoplasty.Reference Onal, Uguz, Kazikdas, Gursoy and Gokce 21 However, other scholars suggested that neither the presence of ear discharge at the time of surgery nor a wet middle-ear mucosa affected the success rate of myringoplasty.
The success rates of graft uptake and hearing improvement were nearly equal in dry and wet ears subjected to myringoplastic underlay.Reference Santosh, Prashanth and Rao 22 – Reference Vijayendra, Rangam and Sangeeta 29 Santosh et al.Reference Santosh, Prashanth and Rao 22 reported that successful graft uptake was evident in 80 per cent of dry and 73.3 per cent of wet ears. Shankar et al.Reference Shankar, Virk, Gupta, Gupta, Bal and Bansal 23 recorded success rates of 88.6 per cent and 80 per cent for dry and wet ears, respectively, after a minimum follow-up time of one year. In addition, Vijayendra et al.Reference Vijayendra, Rangam and Sangeeta 29 considered that the graft take-up rate after tympanoplasty was better when the central perforation was wet rather than totally dry. These results indicate that a wet middle-ear mucosa did not influence the success rate of myringoplasty (type I tympanoplasty).
Regenerative effect of non-growth factors
Interestingly, some studies have suggested that antibiotic ear drops also accelerate traumatic eardrum healing. Lou et al.Reference Lou, Lou, Tang and Xiao 14 found that the topical application of ofloxacin drops in patients with a large traumatic tympanic membrane perforation resulted in an improved closure rate (96.4 vs 65.4 per cent, p = 0.02) and significantly shortened the mean closure time (18.7 ± 6.3 vs 39.1 ± 14.0 days, p < 0.01) compared with the observation group. Yang et al.Reference Yang, Lou, Fu and Zheng 15 reported that the closure rates of the epidermal growth factor and ofloxacin drops groups were similar (93.5 vs 92.0 per cent, p > 0.05). Another study of ofloxacin drops in traumatic tympanic membrane perforation reported closure times of 20.6 days for ofloxacin drops, 24.6 days for ofloxacin drops plus antibiotics, 27.9 days for antibiotics alone, and 35 days for spontaneous healing.Reference Yamazaki, Ishijima and Sato 16
Lou et al.Reference Lou, Lou, Liu and Chang 17 treated moderate and large human traumatic tympanic membrane perforations; ofloxacin drops, basic fibroblast growth factor and Gelfoam patching accelerated closure. One experimental study using ciprofloxacin/dexamethasone and ofloxacin showed that ofloxacin accelerated eardrum healing.Reference Buyten, Kaufman and Ryan 18
A clinical study of cellophane seal and gentamicin ointment observed successful healing rates of 50.0 per cent at 14.7 days with cellophane seal, 86.7 per cent at 18.5 days with gentamicin ointment seal, and 77.8 per cent at 27.7 days with spontaneous healing.Reference Amadasun 19 Gentamicin ointment kept the perforation edge moist and aided eardrum healing. An experimental study showed that topical insulin solution improved the closure rate compared with spontaneous healing (100 vs 80 per cent).Reference Araujo, Murashima, Alves, Jamur and Hyppolito 13
Thus, topical application of ofloxacin drops may be recommended, because it is more easily available and more convenient for otology out-patients than basic fibroblast growth factor or a Gelfoam patch.
Effect of moist environment on regeneration
Can a moist environment accelerate eardrum healing? Does more solution result in faster healing? Lou and colleaguesReference Yang, Lou, Fu and Zheng 15 , Reference Lou, Yang, Tang and Fu 69 reported that in cases of traumatic tympanic membrane perforation it was best to apply ofloxacin drops and epidermal growth factor daily to keep the perforation edge moist, but not wet. Hyaluronic acid retains water and protects the perforation edges from dehydration; most scholars believe that in tympanic membrane perforation, hyaluronic acid mainly keeps the perforation edge moist and aids eardrum healing.Reference Chauvin, Bratton and Parkins 7 , Reference Güneri, Tekin, Yilmaz, Ozkara, Erdağ and Ikiz 70 , Reference Ozturk, Yaman, Cihat, Arbag, Keles and Uyar 71 In an analysis of factors influencing basic fibroblast growth factor in traumatic tympanic membrane perforation, Lou et al.Reference Lou, Yang, Tang and Xiao 8 found that a large dose of basic fibroblast growth factor did not improve the closure rate and prolonged the closure time compared with a lower dose of basic fibroblast growth factor. This might be related to the amount of moisture in the eardrum.
In skin wound healing, excess moisture in the wound bed can impair the healing process and damage the surrounding skin, leading to peri-wound maceration.Reference Okan, Woo, Ayello and Sibbald 39 , Reference Svensjö, Pomahac, Yao, Slama and Eriksson 43 , Reference Vogt, Andree, Breuing, Liu, Slama and Helo 44 If there is excess moisture, healing can be impeded, and there may be subsequent breakdown and deterioration of the wound bed. Studies have also noted greater granulation tissue formation and greater inflammation in moist-treated partial-thickness wounds than in wet-treated wounds.Reference Svensjö, Pomahac, Yao, Slama and Eriksson 43 This suggests that excess moisture (i.e. wet) at the perforation edge adversely affects eardrum healing.
Why do some solutions facilitate eardrum healing and not others? In some comparative studies of saline and other solutions in traumatic tympanic membrane perforation, why did the saline treatment not have a better effect in the same liquid environment?Reference Fina, Baird and Ryan 6 , Reference Chauvin, Bratton and Parkins 7 , Reference Miyoshi, Kawazoe, Igawa, Tabata, Ikada and Suzuki 46 , Reference Mondain and Ryan 50 A study of the effects of embryonic stem cells in traumatic tympanic membrane perforation in Mongolian gerbils found that 100 per cent of tympanic membrane perforations achieved closure with embryonic stem cells after 5 days, compared with only 20 per cent in the physiological saline group.Reference von Unge, Dirckx and Olivius 9 Studies of growth factors and phosphate-buffered saline showed that the closure rate was higher with growth factors than with phosphate-buffered saline.Reference Fina, Baird and Ryan 6 , Reference Chauvin, Bratton and Parkins 7 , Reference Yang, Lou, Fu and Zheng 15 – Reference Lou, Lou, Liu and Chang 17 , Reference Miyoshi, Kawazoe, Igawa, Tabata, Ikada and Suzuki 46 , Reference Mondain and Ryan 50
A recent study showed that eardrum healing was affected by the environmental pH.Reference Akkoc, Celik, Arslan, Demirci, Hucumenoglu and Caydere 20 Normal healthy skin is acidic, and this contributes to the barrier function of skin and decreases bacterial colonisation. An acidic pH inhibits bacterial growth, decreases proteolytic activity, ensures stronger and more regular fibroblasts in vitro, and provides more oxygen support. Most wounds cannot heal spontaneously at an alkaline pH and become chronic. After 7 days, Akkoc et al.Reference Akkoc, Celik, Arslan, Demirci, Hucumenoglu and Caydere 20 found that closure was successful in 5 out of 5 cases at pH 4, in 1 out of 5 cases at pH 7, and in 4 out of 10 cases in the spontaneous healing group. The oedema of the lamina propria and neovascularisation did not differ significantly between pH 4 and 7, while the inflammatory cell infiltration and fibroblastic reaction both differed significantly. The authors proposed that an acidic pH accelerated wound healing by suppressing infection and reducing catabolic and proteolytic activities, or promoted ongoing epithelialisation at the wound edges by inhibiting irregular fibroblast migration to the wound, while an alkaline pH delayed wound healing.
In most experimental studies of growth factors in traumatic tympanic membrane perforation, the vehicle in the control group is phosphate-buffered saline or saline solution,Reference Fina, Baird and Ryan 6 , Reference Chauvin, Bratton and Parkins 7 , Reference Fina, Bresnick, Baird and Ryan 10 – Reference Vrabec, Schwaber, Davidson and Clymer 12 , Reference Mondain and Ryan 50 both of which are alkaline. Gentamicin ointment, ofloxacin drops and insulin solution are acidic. Recent studies have demonstrated that epidermal growth factor can maintain its effect on a chronic tympanic membrane perforation when mixed in a 1 per cent acetic acid solution. This formulation is important for suppressing bacterial overgrowth, particularly of Pseudomonas aeruginosa.Reference Gladstone, Jackler and Varav 5
Based on the evidence, an acidic moist environment at the perforation edges accelerates healing of traumatic tympanic membrane perforations. On the contrary, an alkaline moist environment does not affect the healing process.
Adverse effects of a moist environment
Otorrhoea and external auditory canal discomfort are common adverse effects of tympanic membrane perforation. Most authors suggest that ear drops can introduce bacteria into the external auditory canal, leading to middle-ear infection. However, a few solutions did not increase middle-ear infection when the external auditory canal was cleaned thoroughly before treatment.
There was a dose-dependent increase in the infection rates after basic fibroblast growth factor treatment in rats, and the difference in the infection rates for the groups treated with 10 ng and 100 ng basic fibroblast growth factor was significant relative to a control group (p = 0.03 and p = 0.004, respectively), but not for the group treated with 1 ng basic fibroblast growth factor (p = 0.3).Reference Kälicke, Köller, Frangen, Schlegel, Sprutacz and Printzen 72 Lou and colleaguesReference Lou, Lou, Tang and Xiao 14 , Reference Yang, Lou, Fu and Zheng 15 and Yamazaki et al.Reference Yamazaki, Ishijima and Sato 16 reported that the topical application of ofloxacin drops did not increase the rate of middle-ear infection in traumatic tympanic membrane perforation compared with an observation group.
Surprisingly, two clinical studies on growth factor solutions used to treat human traumatic tympanic membrane perforations found that middle-ear infection did not affect ultimate closure rates.Reference Lou, Lou, Liu and Chang 17 , Reference Lou, Yang, Tang and Fu 69 Some interesting studies even found that acute infection facilitated the healing of chronic tympanic membrane perforation.Reference Derlacki 73 , Reference Spratley, Hellström, Eriksson and Pais-Clemente 74
A study showed that topical ofloxacin drops appeared to increase the incidence of otomycosis.Reference Jackman, Ward, April and Bent 75 However, otomycosis was not reported in clinical studies of ofloxacin drops in traumatic tympanic membrane perforation, perhaps because of the short period of use.Reference Lou, Lou, Tang and Xiao 14 – Reference Lou, Lou, Liu and Chang 17
A few studies showed that excess moisture in the middle ear can induce cholesteatoma of the middle ear.Reference Lou, Tang and Yang 59 , Reference Winerman, Man and Segal 76 A liquid middle-ear environment resulted in oedema of the mucous membrane in the tympanic cavity, aditus ad antrum and Eustachian tube, obstructing gas exchange.Reference Lou, Wang and Su 58 , Reference Lou, Tang and Yang 59 It is possible to induce otitis media effusion in such cases.
Overall, moistening of the perforation edges may allow healing without any complications.
Conclusion and future studies
Moist therapy aids eardrum healing, shortens healing time and does not increase the rate of middle-ear infection. However, excess moisture (a liquid environment) adversely affects healing, substantially prolongs closure time, and increases discomfort and the rate of middle-ear infection. Therefore, for eardrum healing, it is vital to keep the perforation edges moist using appropriate amounts of solution.
Nevertheless, many questions need be studied further, including which agent to use, the best dosage and the best dosage form. The agent should be inexpensive, readily available, and not cause discomfort or ototoxicity. The best dosage and continuous application time of the agent solution need to be determined. The middle ear is connected to the Eustachian tube, and the ear drops could flow out of the Eustachian tube. The external auditory canal anatomy varies, and effective dosages should be evaluated for variants. In addition, it is unclear whether a single dose or regular application is better, and the optimal period of application is unknown. It is important to examine the dosage form. An acidic pH solution may accelerate eardrum healing by suppressing infection. A gel can enable slow release, but is not suitable for eardrops. Keeping moist edges acidic is better for accelerating eardrum healing.
Acknowledgements
This study was supported by the Science and Technology Agency of Zhejiang Province, the Health and Medicine Agency of Zhejiang Province, and the Science and Technology Agency of Yiwu, China (grant numbers 2013C33176, 2015KYB420 and 2015-3-06).