Introduction

An ingrown toenail, or onychocryptosis, is a common problem that occurs when the nail plate punctures the periungual skin causing substantial pain, inflammation, discomfort and increased risk of infection if left untreated [1]. Most cases occur in the hallux and typically present in teenagers and young adults, although may present at any age [2, 3]. Several factors have been proposed as contributory to the occurrence or worsening of ingrown toenails with varying degrees of evidence. These include poor nail cutting technique, hyperhidrosis, ill-fitting footwear, nail deformity, trauma, obesity, and peripheral oedema [4, 5].

Multiple semi-quantitative classification systems have been developed to classify ingrown toenails. Most focus on the severity of ingrown toenails, and generally have three stages: mild (stage I), moderate (stage II) and severe (stage III) [6,7,8]. Although, more recently an alternative approach has been suggested that focusses on the shape of the nail plate and aetiology of the pathology [8]. The performance of these classification systems has not been evaluated but have been proposed as a basis upon which to base treatment decisions.

Conservative approaches in the form of appropriate nail cutting and spicule removal, soaking in warm water, guttering, and orthonyxia (nail bracing) have all been advocated in the literature for use in mild to moderate stages (stage I and II) with varying success rates and quality of evidence [5, 8]. However, when conservative treatment fails, where there is nail deformity, or in more severe cases (stage II and III), a surgical approach is often recommended aiming to remove the problem part of the nail and destroy the underlying matrix to avoid recurrence [8,9,10].

Multiple surgical interventions have previously been described with most including either partial or total avulsion of the nail plate, that is often combined with ablation of the nail matrix to stop regrowth. Nail surgery is performed by a range of health professionals including GPs (general practitioners), orthopaedic surgeons, dermatologists, and podiatrists. Indeed nail surgery forms a substantial part of the workload of podiatrists, having been identified as the tenth most commonly performed procedure performed by the profession [11]. Whilst there is little published data to describe how this common nail pathology is treated in practice, it is clear that a large number of small studies have been published on the topic. Systematically searching for and reviewing these studies, pooling estimates of effectiveness, and providing recommendations for practice and future research is essential to enable evidence-based practice.

A Cochrane review published 10 years ago suggested that use of phenol reduced the likelihood of recurrence but did not differentiate between regrowth of the nail plate (which may be asymptomatic) with recurrence of symptoms [4]. The authors also found that there was insufficient evidence to make recommendations on whether more radical surgery was more effective in cases of more severe disease, or how key patient reported outcomes such as relief of symptoms, patient satisfaction, and post-operative pain were affected by nail surgery. More recent narrative reviews have been written [8, 12], and a systematic review specific to the use of phenol used a very limited search strategy, and did not adhere to key methodological principles such as prospective registration of their review [13]. Accordingly, there is a need for an up-to-date and rigorous systematic review of surgical methods for treating ingrown toenails. The aim of this study, therefore, was to systematically search and synthesise the literature relating to the effectiveness/efficacy of surgical methods for treating ingrown toenails.

Methods

The Cochrane Handbook for Systematic Reviews of Interventions [14] was used to guide the conduct of this review. The review was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [15], This review was prospectively registered at https://www.crd.york.ac.uk/prospero/ [CRD42021251938].

Eligibility criteria

Randomised controlled trials (RCTs) that evaluated the effects of a surgical intervention(s) for ingrown toenails with a follow-up period of at least 1 month were included. Trials comparing one form of surgery with another form of surgery, or a non-surgical intervention, or no intervention were included. Unpublished trials and conference abstracts were only included if the methodological descriptions were adequate to determine eligibility. Where such information was missing from the abstract, it was sought through direct contact with the author. There were no restrictions on the setting, age, or gender of participants. Studies were restricted to English, pertaining to human participants, and must have reported one of the following outcomes for inclusion. Relief of symptoms, and symptomatic regrowth (nail spicules/nail spikes) were considered co-primary outcomes in advance of conducting the searches. Healing time, postoperative complications (e.g., infection and haemorrhage), pain of operation, postoperative pain (duration and intensity) and participant satisfaction were defined as secondary outcomes. After completing searches, and screening it became clear that the majority of papers did not differentiate symptomatic/asymptomatic regrowth of the nail plate, but instead frequently conflated these and considered ‘recurrence’. This was therefore adopted as a co-primary outcome rather than ‘symptomatic regrowth’.

Search strategy

Electronic databases were searched from inception to January 2022: MEDLINE (Ovid), EMBASE (Ovid), CINAHL, Web of Science and Cochrane Central Register of Controlled Trials (CENTRAL). The search strategy (Supplementary File 1) was conducted using Medical Subject Heading (MeSH), truncation and Boolean operators. Other relevant completed and ongoing studies were also sought through screening of clinicaltrials.gov, the International Clinical Trials Registry [ISRCTN] and forward and backwards citations of included studies.

All searches were carried out by the same author and search results generated by the electronic databases were exported to Rayyan, where duplicates were removed. Abstract, titles and full text screening were conducted independently by two review authors, who recorded reasons for exclusion. Discrepancies were discussed with a third author and resolved by consensus. Review authors were not blinded to the author, institution, or the publication source of the study.

Data extraction

A modified Cochrane data extraction form was piloted and then used to extract and record information. Data extracted included: (a) general information such as author(s), title, journal and study funding; (b) trial characteristics such as study aim and objectives, study design, unit of allocation and ethical approvals; (c) participant characteristics such as setting, inclusion/exclusion criteria, sample size (number of participants and nail folds), age, gender, baseline imbalances, severity of ingrown toenails; (d) intervention and comparison group(s); (e) outcome measures including as time points, unit of measurement, outcome definition, data at baseline/follow-up and statistical methods.

Two review authors independently extracted data from the included studies, with disagreements resolved through consensus of a third review author. Where data were missing or unclear, the corresponding author(s) was contacted via email and relevant information requested. If after initial request no response was forthcoming, at least one further email was sent.

Data synthesis and analysis

Meta-analyses were conducted using Review Manager (RevMan V5.3 Cochrane Collaboration, Oxford, UK). The RevMan programme is designed to allow independent data entry by different reviewers. This ensures blinded data entry and highlights any discrepancies in values entered by reviewers. Any discrepancies between entered values was rechecked and discussed. For continuous data, pooled results were expressed as mean differences (MD), 95% confidence intervals (CI) and p values, with <0.05 considered significant. For dichotomous outcomes, pooled results were expressed as risk ratios (RR) and corresponding 95% CI, generated by RevMan. Where studies used a different assessment tool to measure the same construct, the standardised mean difference (SMD) and corresponding 95% CI was calculated. Where studies had multiple assessment time points, data were extracted for the final follow-up time point from randomisation.

Statistical heterogeneity and consistency were determined by interpreting the I2 statistic, and the following thresholds were identified a priori: 0–40% may not be important, 30–60% may represent moderate heterogeneity, 50–90% may represent substantial heterogeneity, 75–100% considerable heterogeneity [14]. If statistical heterogeneity was noted (I2 > 40%) for a particular treatment comparison, a random-effects model was used for analysis to account for expected heterogeneity between studies. Where there was no or little evidence of statistical heterogeneity, a fixed-effects model was used. The fixed-effect model assumes all studies are measuring the same treatment effect and that all differences between studies are due to random (sampling) error. The Mantel-Haenszel methods [16] is the default fixed-effect method implemented in Revman. The random-effects model assumes the treatment effect varies between the studies. This model estimates the mean of the distribution of effects and is weighted for both within-study and between-study variation. This approach uses the variance within each study and adds a second measure known as Tau2. Sensitivity analysis was performed if a substantial heterogeneity (I2 > 75%) was detected. Where data aggregation was not possible due to methodological heterogeneity, these results were summarised narratively. Potential publication bias for each outcome with more than ten studies was evaluated by visual inspection of funnel plots [17].

Certainty of evidence

The Grades of Research, Assessment, Development and Evaluation (GRADE) approach was used independently by two review authors to analyse the certainty of evidence against six domains: risk of bias/ certainty, indirectness of evidence, heterogeneity or inconsistency of effect, imprecision and publication bias [18, 19]. Disagreements were resolved by a third author.

Risk of bias

The Cochrane Risk of Bias tool (RoB 2.0) was assessed independently by two review authors to determine the validity and methodological rigor [18]. Discussion between the two review authors was utilised to resolve any discrepancies, with any disagreements resolved by a third author. Included studies were assessed on the randomisation process, deviations from the intended interventions, missing outcome data, measurement of the outcome and selection of the reported result, with each domain judged as ‘low’, ‘some concerns’, or ‘high’ risk. The overall RoB judgement was derived from the highest classified domain.

In this paper we present analysis for our primary outcomes and secondary outcomes are presented in a subsequent paper.

Results

Search results

A total of 3,928 records were identified, with 1,641 remaining after de-duplication, 70 studies were retrieved and examined for full-text screening leaving 36 studies eligible for inclusion (Fig. 1 and Supplementary Table 1). Of these, 31 were included in the meta-analysis for recurrence and 5 were reported narratively. In accordance with Cochrane Handbook guidance [Section 23.3], the study by Tatlican and colleagues [20] was not included in the meta-analysis for the following reasons: 1) groups could not be combined due to the similarity in groups, comparison of the same chemical at different application timepoints, 2) all groups were relevant therefore it was not possible to include only relevant groups, and 3) as all groups were being compared there was no single comparator to spilt the ‘shared’ group into two or more groups. Other reasons for exclusion from the meta-analysis included not assessing the outcome measure [21], heterogeneity in terminology reporting [22,23,24] and three trial arms [20].

Fig. 1
figure 1

PRISMA flow diagram of literature search and study selection phases. n, number; WoS, Web Of Science; CENTRAL, Cochrane Central Register of Controlled Trials; WHO ICTRP, World Health Organisation International Clinical Trials Registry Platform; ISRCTN, International Standard Randomised Controlled Trial Number Registry

Full size image

In addition, synthesis of evidence for relief of symptoms was reported narratively due to the small number of studies reporting this outcome.

Study characteristics

Characteristics of included studies are summarised in Table 1. Of the 36 RCTs, published between 1979 to 2021, studies were conducted across 14 countries with Turkey being the most frequent (16.6%). Included studies comprised of 3,756 participants, with a sample size ranging from 10 to 125 per study. Five studies [22,23,24,25,26] did not report gender. Of the remaining 31 studies, 62.7% of participants were male.

Table 1 Characteristics of included studies
Full size table

Interventions

Of the 36 included studies, 5 compared a conservative intervention: orthonyxia [46], nail tube splinting [29], nail elevation and flexible tubing [36], or gutter method [24, 52] to either chemical or surgical matrixectomy. Eleven studies [26, 28, 35, 38, 41, 45, 47, 49, 50, 53, 55] compared chemical matrixectomy to surgical matrixectomy, using various techniques. Two studies [39, 43] compared chemical matrixectomy to avulsion only, and two studies [31, 41] compared chemical matrixectomy to a combination of surgical and chemical matrixectomy.

One study compared a surgical intervention to another surgical intervention [40], two added the use of phenol to a surgical intervention [32, 41]. Three studies compared the same surgical intervention, one adding phenol alongside partial nail avulsion [43], another compared nail avulsion with phenol or trichloroacetic acid [33] and one introduced a new suturing technique alongside a Winograd procedure [54]. One study compared chemical matrixectomy with phenol or trichloroacetic acid [27]. Four studies compared the same surgical intervention and introduced a chemical matrixectomy at different application timings [20, 22, 23, 51].

Five compared an alternative intervention: CO2 laser [42] or electrocautery [25, 34, 44, 48] to either chemical or surgical matrixectomy. There were 2 studies [30, 37] comparing local anaesthetics (4 mL solution of 2% mepivacaine; 2% plain lidocaine, respectively), with or without epinephrine and 2 studies [21, 35] looked at pre- and postoperative use of antibiotics following a surgical intervention.

Recurrence

Recurrence was reported in all but one study [21]. The definition of recurrence varied between studies (Table 2) and two studies were unclear and reported the ‘number of successes/number of failures’ [24] and ‘number symptom-free’ [22, 23], respectively. Therefore, these studies were reported narratively. Follow-up ranged from 1 to 24 months.

Table 2 Recurrence definitions
Full size table

Chemical matrixectomy vs conservative management

Two studies [29, 36] found that phenol matrixectomy did not significantly decrease the rate of recurrence when compared to a conservative approach such as nail tube splinting or nail elevation and flexible tubing (RR 0.55 [95% CI 0.19 to 1.61], I2 0%; p = 0.280) (Fig. 2).

Fig. 2
figure 2

Meta-analysis comparing A Chemical matrixectomy vs Conservative management and B Surgical matrixectomy vs Conservative management

Full size image

Surgical matrixectomy vs conservative management

Two studies [46, 52] compared surgical matrixectomy to a conservative approach, however neither method was significantly more effective at preventing recurrence (RR 0.72 [95% CI 0.33 to 1.56] I2 34%; p = 0.410) (Fig. 2). Conversely, one study [24] that could not be included in this meta-analysis reported the wedge resection (27/32, 84%) to be superior to the gutter treatment (20/36, 56%) in terms of ‘number of successes’ (p < 0.05).

Chemical matrixectomy vs surgical matrixectomy

Combining the eleven studies [26, 28, 35, 38, 41, 45, 47, 49, 50, 53, 55] where chemical matrixectomy was compared to surgical matrixectomy, found no significant difference in their ability to prevent recurrence (RR 0.75 [95% CI 0.46 to 1.21] I2 55%; p = 0.230) (Fig. 3). In addition, funnel plots suggest an absence of publication bias (Supplementary File 2) as these data are symmetrically distributed.

Fig. 3
figure 3

Meta-analysis comparing Chemical matrixectomy vs Surgical matrixectomy

Full size image

Chemical matrixectomy vs chemical matrixectomy

Two studies [27, 33] compared phenol to trichloroacetic acid, however neither chemical proved to be more effective at preventing recurrence (RR 0.19 [95% CI 0.01 to 3.80] p = 0.280) (Supplementary Fig. 1).

Surgical matrixectomy vs other (e.g., CO2 laser, electrocautery)

In three studies [34, 42, 44], a surgical intervention (partial nail matrixectomy or curettage) did not significantly decrease the rate of recurrence when compared to an alternative method of matrixectomy such as electrocautery or CO2 laser (RR 1.61 [95% CI 0.88 to 2.95] I2 37%; p = 0.120) (Supplementary Fig. 2).

Chemical matrixectomy vs other (e.g., CO2 laser, electrocautery)

Similarly, when comparing chemical matrixectomy to an alternative method of matrixectomy in two studies [25, 48], there was no significant difference in prevention of recurrence (RR 0.58 [95% CI 0.25 to 1.37] I2 0%; p = 0.220) (Supplementary Fig. 3).

Avulsion vs avulsion + chemical matrixectomy

Avulsion with phenol matrixectomy was compared with nail avulsion alone in two studies [39, 43]. There was a significant reduction of recurrence in favour of phenol matrixectomy (RR 0.13 [95% CI 0.06 to 0.27] I2 0%; p < 0.001) (Supplementary Fig. 4). Although Greig and colleagues [39] called their procedure ‘Nail edge excision’ the procedure described was the same as partial avulsion.

Surgical matrixectomy vs surgical matrixectomy

One study [54], compared the Winograd procedure using a new suturing technique, compared to the same surgical intervention and a traditional suturing technique. After 12 months, participants were asked to report any recurrence via telephone. The new suturing technique was more effective at preventing recurrence compared to the traditional technique (RR 0.42 [95% CI 0.21 to 0.85]) (Supplementary Fig. 5).

Another compared central toenail resection to wedge toenail resection [40]. After 6 months, the central toenail resection was considered more effective at preventing recurrence compared with the wedge toenail resection (RR 0.05 [95% CI 0.0 to 0.79]) (Supplementary Fig. 6).

Surgical matrixectomy vs surgical + chemical matrixectomy

Two studies compared a surgical intervention, either nail bed excision or wedge resection, with the same surgical intervention plus the addition of phenol [32, 41]. However, addition of phenol was not significantly more effective at preventing recurrence (RR 3.68 [95% CI 0.20 to 67.35] I2 76%; p = 0.380) (Supplementary Fig. 7).

Chemical matrixectomy vs surgical + chemical matrixectomy

In two studies [31, 41], surgical matrixectomy plus phenolisation did not significantly decrease the rate of recurrence when compared to phenolisation alone (RR 1.92 [95% CI 0.06 to 62.30] I2 62%; p = 0.710) (Supplementary Fig. 8).

Local anaesthetic vs local anaesthetic + adrenaline (epinephrine)

Two studies [30, 37] compared local anaesthetic (4 mL solution of 2% mepivacaine; 2% lidocaine, respectively), with a combination of the same local anaesthetic plus adrenaline (epinephrine). The use of adrenaline did not significantly decrease the rate of recurrence (RR 1.03 [95% CI 0.22 to 4.86] I2 0%; p = 0.970) (Supplementary Fig. 9).

Chemical application time: 30 vs 60 s

Four studies used the same surgical intervention but varied the duration that the chemical was applied during the matrixectomy [20, 22, 23, 51]. Of these, three studies [20, 22, 23] could not be included in the meta-analysis, and none reported significant differences in chemical timing applications. The study by Gem and colleagues [22] compared chemical ablation with either 3 min of 80% phenol or 2 min of 10% sodium hydroxide and the second [23] compared either 1 or 2 min with 10% sodium hydroxide. Of the 422 procedures, 148 were lost to follow up, leaving 248/274 (study 1 n = 140/157; study 2 n = 108/118) who were completely asymptomatic at 18 months. No significant differences were found between the interventions. The numerical data was reproduced faithfully from the publication. There is an arithmetical error, but this has not been corrected due to uncertainty where it occurs. Lastly, Tatlican and colleagues [20] compared phenol with partial nail avulsion at 1, 2 and 3 min on rates of recurrence, assessed every 6 months over 24 months, and found no significant difference between the three groups (p = 0.092).

Of the one study [51] that was included in the meta-analysis, Muriel-Sanchez and colleagues compared the recurrence rate between phenol applications of 30 or 60 s, finding the 60-s application was more effective at preventing recurrence compared to the 30-s phenol application (RR 2.00 [95% CI 0.19 to 21.41]) (Supplementary Fig. 10).

Antibiotics vs no antibiotics

Bos and colleagues explored the use of topical antibiotics (5.3 mg soluble tablet of gentamicin applied locally) on recurrence, with and without matrix excision and phenol [35]. After 12 months the use of topical antibiotics alongside a chemical or surgical matrixectomy did not significantly decrease the rate of recurrence (RR 0.54 [95% CI 0.12 to 2.52] I2 58%; p = 0.430) (Supplementary Fig. 11).

Relief of symptoms

Five studies assessed relief of symptoms [22, 23, 38, 40, 46]. Two studies assessed symptoms using a visual analogue scale ranging from 0 to 10 [38, 46], the remaining three studies did not specify the instrument used [22, 23, 40]. No definitions were provided for relief of symptoms.

Chemical procedures

In two studies [22, 23], no statistically significant differences were identified between patients receiving 3 min application of 80% phenol, 2 min of 10% sodium hydroxide and 1 min of 10% sodium hydroxide. However, Gem and colleagues did report 91% of all participants were asymptomatic after a minimum follow up time of 12 months (study 1) [22] and 18 months (study 2) [23].

Chemical and surgical procedures

Despite a tendency in the matrix group to have fewer persisting symptoms, the study by Gerritsma-Bleeker and colleagues [38] found no significant differences between partial nail extraction with phenolisation and partial nail extraction with matrix excision at 1, 3 or 12 months (p = 0.130, p = 0.270, p = 0.290, respectively).

Surgical procedures

Habeeb and colleagues [40] showed central toenail resection was significantly better in relieving symptoms compared to wedge toenail resection after 4 and 8 weeks (both p = 0.001).

Surgical and conservative procedures

Following receipt of either partial nail extraction with partial matrix excision or orthonyxia, no differences were noted in Kruijff and colleagues [46] study after 12 months.

Ongoing studies

One ongoing clinical trial (CTRI/2017/09/009951) of interest was identified. Registered in 2017, this study remains classified as ‘Not yet recruiting’. Attempts were made to obtain an update on progress from the listed investigators but with no success. Three trial registries of interest were also identified (NCT03732313; IRCT201604176403N6; ACTRN12619001719123), however results were already included in this review [40, 42, 51].

Risk of bias

We used the used the Cochrane RoB 2.0 tool and assessed six domains for each study. No study was rated as low risk, for reasons such as not or providing information surrounding the randomisation process, not including all participants in the final analysis and failing to provide information on blinding of participants or the outcome assessor. Risk of bias summaries are presented in Fig. 4 and risk of bias table in Supplementary Table 2.

Fig. 4
figure 4

Risk of Bias Summary Plot: RoB 2.0 Tool

Full size image

Certainty of evidence

The certainty of evidence (Supplementary Table 3) for the outcome recurrence was: very low for meta-analyses comparing surgical vs conservative (2 RCTs, n = 209), chemical vs surgical (11 RCTs, n = 1041), surgical vs other (3 RCTs, n = 388), chemical vs avulsion (2 RCTs, n = 263), surgical vs surgical + chemical (2 RCTs, n = 171), chemical vs other (2 RCTs, n = 160), chemical vs surgical + chemical (2 RCTs, n = 191), epinephrine vs without epinephrine (2 RCTs, n = 114). Low for chemical vs conservative (2 RCTs, n = 173), phenol vs trichloroacetic acid (2 RCTs, n = 187) and surgical vs surgical + suturing (1 RCT, n = 128). Moderate for surgical vs surgical (1 RCT, n = 100), chemical timings (1 RCT, n = 108) and antibiotics (1 RCT, n = 117). The main reasons for downgrading the evidence were risk of bias, indirectness of evidence and imprecision.

Discussion

Ingrown toenails, or onychocryptosis is a common nail pathology. Surgical resolution is often sought when conservative measures fail, or even as a first line intervention in more severe cases. As such, nail surgery is one of the most commonly performed procedures by podiatrists in the UK [11]. Despite the high number of publications on the topic, there has been a lack of robust systematic reviews covering the spectrum of surgical options in the decade since the last Cochrane review [4].

This review followed Cochrane methodology and conducted a prospectively registered systematic review with meta-analysis of surgical treatments for ingrown toenails. This paper includes a detailed description of our methodology and presents findings from our predefined primary outcomes: recurrence and relief of symptoms. Analysis of secondary outcomes will follow in a subsequent publication.

The systematic, search identified 1,641 potential publications which, after screening, enabled 36 studies with 3,756 participants covering a range of techniques that were included in the review. This is a substantial increase on the 24 and 18 studies in the previous Cochrane review [4], and the review by Vinay and colleagues [13], respectively. Recurrence was reported in all but one study, although there were variations in how this was defined and captured. Meta-analysis did not demonstrate a difference in risk of regrowth for most comparisons. Perhaps unsurprisingly, phenolisation was a notable exception to this pattern and when compared to nail avulsion alone, there was a very low certainty of evidence that use of phenol significantly reduced the risk of recurrence [39, 43]. However, use of phenol combined with surgical excision offered no benefit over phenolisation alone [31, 41]. In terms of how long to apply the phenol for, there was a moderate certainty of evidence that application of 1 min had lower risk of regrowth compared to 30 s [51], but there was no additional benefit when it was applied for 2 or 3 min [20]. Studies of peri-operative factors beyond the actual procedure such as use of different local anaesthetic with / without adrenaline [30, 37] and topical antibiotics [35] did not affect rates of regrowth even with such an atypical application technique.

Surprisingly, symptom relief was only reported in five [22, 23, 38, 40, 46] of the 36 studies and in three of those, it was not clear whether this was patient reported, or determined by the clinicians [22, 23, 40]. Even in these studies, exactly what ‘symptoms’ refers to is often unclear. Ingrown toenails are intensely painful though, and that this is rarely captured is a poor reflection on the quality of research in the field: it is no longer acceptable for studies to fail to capture key outcomes that matter to patients, and instead only focus on clinician reported outcomes. The importance of PROMs is well recognised by major national health policy and regulatory authorities [56, 57]. The authors question whether it is acceptable for future trials in ingrown toenails to continue to omit patient reported outcomes. It is important that future clinical trials differentiate between regrowth, which may be asymptomatic, and regrowth which causes pain and infection.

Clinical conclusions from this paper should be interpreted in line with our second paper that considers the secondary outcomes from our review: healing time, postoperative complications, pain of operation, postoperative pain (duration and intensity), and participant satisfaction. Only with these can a broader, more holistic, assessment of outcome be fully appreciated so these are essential for guiding practice.

All 36 studies included in the review were assessed as being either high risk or having some concerns about bias when assessed with the Cochrane RoB 2.0 tool. Similarly, out of the 15 comparisons made here, most were considered to have either very low, or low certainty of evidence when assessed with the GRADE system. Only three reached moderate, and none were considered to have high certainty. The main reasons for downgrading the evidence were risk of bias, indirectness of evidence, and imprecision. In addition, surgical technique was often poorly described, and there was large variation in the use of terms such as recurrence. It is also important that future clinical trials differentiate between regrowth that may be asymptomatic and regrowth that causes pain and infection. To put this another way, 3,756 people have taken part in research studies that do little to guide clinical practice. Some of this may be due to poor reporting, but poor design also plays a major role. Regardless, both of these reasons can, and should be avoided and this topic has been widely discussed in the literature with recommendations made to improve research across healthcare [58,59,60]. Findings from this review differ from those of previous reviews. In part, this may be explained by the publication of new research in the decade since the Cochrane review [4] and a broader focus than the review by Vinay and colleagues specific to phenol [13]. However, the risk of bias assessment was also different as was the grading of strength of recommendations that could be made. Whilst there would inevitably be some variation in these relatively subjective assessment systems the authors stand by this assessment and have discussed some of the methodological limitations in the existing evidence base that have led us to this conclusion. More, high quality clinical trials to inform clinical decision making are urgently needed in nail surgery.

This review and meta-analysis both have strengths and limitations. The authors consider the robust methodology of the search, screening, extracted data, synthesis, meta-analysis, and use of tools such as RoB 2.0 and GRADE as methodological strengths. Deliberate attempts have been made to ensure that comparisons within the meta-analysis are clinically meaningful. Whilst some readers may disagree with how these studies have been compared, or want additional comparisons, they have been made in an open and transparent way. As a further note, this process was made more difficult due to the poor procedure descriptions, with many describing more than one procedure i.e., stated as nail edge excision but partial avulsion was described. Well established reporting guidelines such as the SUPER and IDEAL frameworks should be followed in the future [61, 62].

Conclusion

This paper presents the co-primary outcomes from a systematic review with meta-analysis that should be interpreted in conjunction with its second paper. Despite the high number of publications on this topic, the quality of research was poor and the conclusions that can be inferred from existing trials is limited. Phenolisation of the nail matrix reduces the risk of recurrence following nail ablation, and 1 min appears to be the optimum time for application but there is less certainty around this recommendation. Further research is needed to explore the effectiveness of other commonly used ablative agents such as sodium hydroxide and to systematically explore the optimisation of post-operative care.