Wound healing in pre‐tibial injuries – an observation study

Abstract

Pre‐tibial lacerations are complex wounds affecting a primarily aged population, with poor healing and a potentially significant impact on social well‐being. Management of these wounds has changed little in 20 years, despite significant advances in wound care. A retrospective observational study was undertaken to observe current wound care practice and to assess the effect of various medical factors on wound healing time on 24 elderly patients throughout their wound journey. Wound length was found to be substantively and significantly associated with wound healing time, with a reduction in instantaneous healing rate of about 30% for every increase of 1 cm in wound length. Hence, longer wounds are associated with longer wound healing times. Prescription of several categories of drugs, including those for ischaemic heart disease (IHD), hypertension, respiratory disease or asthma; and the age of the patient were not significantly associated with wound healing times, although substantive significance could be inferred in the case of prescription for IHD and asthma. Despite the small sample size, this study identified a clear association between healing and length of wound. Neither the comorbidities nor prescriptions explored showed any significant association although some seem to be more prevalent in this patient group. The study also highlighted other issues that require further exploration including the social and economic impact of these wounds.

INTRODUCTION

Lacerations or wounds to the pre‐tibial area on the lower leg have long been identified as being some of the most serious wounds presenting to emergency departments (EDs), due to the incidence, prolonged healing (or non healing) time and an often marked impact on the social welfare of patients 1, 2.

The majority of patients sustaining pre‐tibial lacerations that require hospital treatment are elderly, with a number of causes underlying this phenomenon. The most significant is likely due to structural and physiological changes in the skin, which render it firstly more susceptible to damage, and secondly less able to heal and repair following trauma. A flattening of the dermal–epidermal junction renders the skin more susceptible to shearing forces, which can cause separation of the skin layer (3). The development of ‘thin’ skin can be anecdotally observed in many older people, with Richey et al. (4) detailing a marked ‘thinning’ in those over the age of 70, and especially in females.

In addition, elderly people are more likely to experience other underlying or chronic disease states, or be taking medication which can either impair the healing process or increase the likelihood of traumatic incidents in the first place. For instance, many of the older population may be prescribed aspirin or warfarin, which will affect bleeding and clot formation, or they may be taking steroids which may impede healing (5). Equally, poor eyesight or locomotive instability will increase the chance of tripping or striking sharp edges and sustaining injuries.

Pre‐tibial wounds may be treated either surgically, with debridement and skin flap creation, or conservatively, with or without adhesive strips to appose wound edges 5, 6. Other wound closure techniques such as suturing or stapling may cause more trauma to the already fragile skin, increasing necrosis and healing time (7). Ongoing wound management is generally undertaken by community services following presentation and initial wound care; so little is known or learned about the outcome of these wounds by the ED staff who have treated them. Locally, conservative treatment with adhesive strips has been challenged by community staff who identified that the exudate from wounds, which is common, can cause adhesive strips to lift off. This can leave the wound edges unapposed with a mass of adhesive strips proximal to the wound bed, adding nothing except to provide an environment ripe for bacterial contamination. Pre‐tibial injuries are associated with increased care needs, and dependence, although it is unclear if this is a direct result of the injury, or if the injury was a reflection of general health decline (2).

Research into the closure of the wounds has concentrated mainly on the efficacy of surgical closure techniques, even though it is widely accepted, anecdotally, that only a small number of injuries are ever referred for surgical management. This study aimed to observe wound healing, to describe the population and the wounds sustained, and to identify factors that may correlate with healing/non healing.

METHODS

The study was undertaken in a district general organisation which has two EDs, each treating approximately 70 000 new patients every year. Data was also retrieved by examination of the ED electronic information system (EDIS), on which all clinical/clerical information is stored. Details included patient demographics and wound descriptions. Data relating to ongoing care was collected via telephone consultation with the community teams involved. Ethics approval for the study was granted through the University of Huddersfield, and Research and Development approval granted by the NHS organisations involved.

All patients, over 60 years old, presenting between 1 October and 31 December 2009, were included in the study and followed up in the community until the point of healing or for a period of 120 days (representative of classification as chronic wound).

Statistical analysis

The wound healing analysis was formulated in a time‐to‐event analysis framework, in which the time to the event of wound healing was considered. Hence, a positive outcome was represented by the occurrence of the event of wound healing, whereas a negative outcome was represented by a censored observation, in which there was no evidence of wound healing. Censored observations included those whose wounds had failed to heal by the end of the study at 120 days. No participants were lost to follow‐up during the course of the study.

Two series of analyses were conducted. Firstly, the substantive effect of categorical factors on wound healing time was assessed in a series of univariate analyses using the Kaplan–Meier method. Initially this analysis was performed on the full sample. The data was then partitioned by each of the following categorical factors in turn: whether or not the patient was receiving a prescription for each of ischaemic heart disease (IHD), hypertension, chronic obstructive pulmonary disease (COPD) or asthma. The statistical significance of each factor considered on wound healing time was assessed by the calculation of log‐rank statistics. Survival curves were constructed indicating the proportion of wounds remaining unhealed at a given time. Median wound healing times, with associated 95% confidence intervals (CI) where appropriate, were also derived.

A predictive multivariate analysis was then undertaken using the Cox semi‐parametric model to assess the significance of all categorical factors considered above, plus continuous covariates corresponding to age and wound length.

RESULTS

Over the 3‐month study period data on 24 patients was reviewed. Most of the patients were female (95·8 %, n = 23), with a mean age of 82. A high proportion of the total population were in the 85–90 year age group (62·5%). The study sample presented with multiple medical conditions; the most common of which were IHD (25%), stroke (29·2%), high blood pressure (25%), COPD (29·2%) and asthma (25%). The most common prescribed medications or drug groups documented in the study population were aspirin (45·8%), diuretics (45·8%), statins (37·5%), antidepressants (29·1%) and salbutamol (25%).

Wound length ranged from 2 to 11 cm, with a mean wound length of 5·9 cm. Of the 24 patients enrolled, 16 wounds healed within the first 90 days. No further wounds were recorded as healing by the end of the study after 120 days; hence, eight patients were identified as non healers. For the full sample, the median wound healing time was found to be 41·0 days (CI, 7·4, 74·6).

The corresponding wound survival curve is shown in Figure 1. Censored observations are indicated by marks on the survival curve.

Figure 1.

Wound survival curve for full data set. Proportion of wounds healing as a fraction against healing time in days. Censored data marked as + points (not all visible because of overlap).

Table 1 shows the median wound healing times, plus 95% CI were calculated, for the data set partitioned by all categorical factors in turn. CI for wound healing times for those patients with prescriptions for hypertension and asthma could not be calculated with any accuracy, because of an insufficient number of non censored observations in the sample. Log‐rank statistics and corresponding significance values are also presented.

Table 1.

Median wound healing survival times and confidence intervals for partitioned data: univariate models

Factor	Wound healing time (days) factor present		Wound healing time (days) factor absent		Log‐rank statistics	Significance
	Median	95% CI	Median	95% CI
Prescription for IHD	27·0	(12·6, 14·1)	55·0	(0, 115·3)	1·079	0·299
Prescription for hypertension	41·0	–	31·0	(0, 78·8)	0·537	0·464
Prescription for COPD	55·0	(0, 147·4)	41·0	(10·1, 71·9)	0·087	0·769
Prescription for asthma	74·0	–	27·0	(20·1, 33·9)	1·302	0·254

CI, confidence intervals; IHD, ischaemic heart disease; COPD, chronic obstructive pulmonary disease.

The corresponding wound survival curves are shown in 2, 3, 4, 5 Censored observations are indicated by marks on the survival curve (some are overlapping, so only four are visible).

Figure 2.

Wound survival curves for patients with and without ischaemic heart disease (IHD) prescription. Proportion of wounds remaining unhealed against healing time in days.

Figure 3.

Wound survival curves for patients with and without hypertension prescription. Proportion of wounds remaining unhealed against healing time in days.

Figure 4.

Wound survival curves for patients with and without chronic obstructive pulmonary disease (COPD) prescription. Proportion of wounds remaining unhealed against healing time in days.

Figure 5.

Wound survival curves for patients with and without asthma prescription. Proportion of wounds remaining unhealed against healing time in days.

Table 2 shows the P‐values, parameter coefficients with standard errors, odds ratios and associated 95% CI for each of the factors and covariates considered in a multivariate Cox analysis. For categorical factors, the odds ratio corresponds to the hazard ratio for a patient in a particular category compared to the reference category; where the hazard ratio (relative hazard) is the ratio of instantaneous healing rates of wounds in the two categories. For continuous covariates, the odds ratio corresponds to the hazard ratio when the value of the covariate is increased by 1 unit.

Table 2.

Median wound healing survival times and confidence intervals: multivariate Cox model

Covariate	P‐value	Parameter coefficient (SE)	Odds ratio	95% CI
Prescription for IHD
No prescription (reference)	0·797	−0·178 (0·692)	0·84	(0·22, 3·25)
Prescription
Prescription for hypertension
No prescription (reference)	0·434	0·656 (0·838)	1·93	(0·37, 10·0)
Prescription
Prescription for COPD
No prescription (reference)	0·464	0·686 (0·936)	1·99	(0·32, 12·4)
Prescription
Prescription for asthma
No prescription (reference)	0·374	−1·033 (1·162)	0·36	(0·04, 3·47)
Prescription
Wound length	0·007	−0·372 (0·139)	0·69	(0·52, 0·91)
Age	0·514	−0·027 (0·041)	0·97	(0·90, 1·06)

CI, confidence intervals; IHD, ischaemic heart disease; COPD, chronic obstructive pulmonary disease.

All but one of the wounds was closed with adhesive strips, and none was referred for surgical debridement. Only half the patients (50%) had documented evidence of non adherent dressing being used in the ED, in contrast to 79·2% in the community. In contrast to the ED, where a limited range of dressings were used, documentation in the community showed a variety of products being used with limited evidence of rationale for their use. Although bandaging is mentioned in the community treatment, there is no evidence that this is compression bandaging.

In terms of follow‐up, all patients had their ongoing care in the community, 79·2% by district nurses and 20·8% by their practice nurse, with visits ranging from one to four times each week. Three patients were reviewed in the ED, but this was only for one clinical appointment prior to referral to community services. During data collection, PU identified that a number of patients had an increase in their care needs following injury, with either admission to residential care, or an instigation of, or increase in, care visits to their home. This was not formally tested or included in the analysis.

DISCUSSION

Pre‐tibial lacerations are a relatively common injury, affecting mainly elderly women, which is again reflected in this study. Reports of injury prevalence range from 0·4 to 0·7 per 1000 admissions (1), and 5·2 per 1000 in a UK study (8), although this latter figure was an estimate of attendance made by A&E staff, rather than diagnostic code analysis. Our study identified 24 patients over a 3‐month period, which would translate to a prevalence of approximately 0·6 per 1000 attendances; a figure which, although lower than expected, correlates well with the results of Laing et al. (1). Speculation on why elderly women continue to show higher incidence of injury has been accounted for by their increased life expectancy, and even to their attire – dressing in skirts, with little or no protection of the pre‐tibial area. Laing et al. (9) studied this effect and identified a reduction in force being transmitted through multiple, or thicker, layers of clothing.

The prevalence of concurrent medical conditions and medication is not studied in the pre‐tibial literature but holds few surprises. It was hypothesised by the team that patients with pre‐tibial injuries would most likely suffer from chronic conditions that might, especially through treatment, lead to poor skin integrity and healing. Nevertheless, when compared to national data, where it was available, some conditions and medications were clearly more common. Twenty five percent of the study group had IHD, against a national prevalence in the UK of only 3·6%, whilst 29·2% had sustained a stroke, which is grossly different than the national figure of 1·7% (10). The prevalence of other conditions traditionally associated with poor healing, for example, diabetes and heart failure, was low in this study; so, it was not possible to draw any statistical evidence of association. Meanwhile, it is not surprising that aspirin, statins and diuretic treatments are highly common considering study group age and the prevalence of cardiac/vascular disease. It is of interest that many were prescribed antidepressants (29·1%) or benzodiazepines (16·7%), although the national prevalence of these drugs in this population is unknown. Prescriptions for IHD or asthma appeared to be substantively associated with wound healing time, with longer healing times associated with no prescription for IHD, and with prescription for asthma. However, no statistical significance was seen and a large degree of uncertainty in median healing times may be observed (Table 1). Considering the sample as a whole, there seemed to be a rapid rate of wound healing observed up to about 25 days, which then slows markedly over the period of 25–75 days, with no further healing events after 75 days (Figure 1). Differences in the proportion of wounds remaining unhealed are more apparent in all cases towards the latter part of the study.

The effect of wound length appears more unambiguous. Table 2 shows that it is the only substantive and significant predictor of time to wound healing. All wounds less than 7 cm in length healed within 70 days, whereas only a minority of wounds of length 7 cm or more had healed by 120 days. The odds ratio of 0·69 indicates that the ‘hazard’ of wound healing (instantaneous healing rate) reduces by 31% for every 1 cm of additional wound length, controlling for other variables. No other study identifies wound length as a predictor of outcome. A larger sample size, with more robust description or classification of wounds, may provide better understanding of the effects of wound length, depth and shape on healing.

As was expected, most of the wounds were closed with adhesive strips, with only one wound being sutured. This correlates with the literature and a general consensus that adhesive strips are the least traumatic method of wound closure 5, 6, 7. Adhesive strips apply less tension to the typically friable skin, whereas sutures would be more likely to further damage skin edges. In wounds with skin flaps, increased rates of skin edge necrosis and slower wound healing rates have been observed when suturing was carried (7) out. Adhesive strips do have their drawbacks, however, as adhesion relies on contact with clean, dry skin which may be compromised with the presence of exudate or with a moist dressing. Equally, healing may be slower than rates described in the surgical closure papers, and because none of the study sample was referred for surgical intervention, it is not possible to identify differences in healing times. Early surgical intervention may be more cost‐effective for a proportion of patients, where delayed healing will affect long‐term health and social needs. It is not clear if the lack of on‐site specialist Plastics services affected the referral to specialty care for this population.

One element of management that was noted by the study lead was the variation in practice in wound management across the community teams. Although there was some variation in the A&E department care, there is a relatively limited availability of dressing options. In contrast, documentation from the community notes showed a wide range of dressings used, with some evidence of duplication even within one dressing episode, for example, two non adherent dressings applied at the same time. This may relate to the autonomous nature of community work, but happens despite both acute and community teams working from a shared wound care formulary. Although there is some mention of bandaging, it is not clear if this was compression bandaging, which may be useful in the promotion of healing, but is poorly discussed in the pre‐tibial literature 6, 11. Standardisation of practice could lead to efficiencies in services and potentially a reduction in disturbance of the wound.

Other data of note that may warrant greater investigation is the number of visits/contacts made throughout the wound journey, and the increase in dependency of the patients across this time. Both of these issues account for significant resource implications for health and social care, and require further investigation.

The most significant limitation of the study was the sample size, affecting analysis of much of the data, providing the opportunity only to undertake a survival analysis using a limited number of variables. As a pilot study, it has provided some insight into the key variables which require further exploration, including the links to healing, patient dependency and variable wound care practices across primary and secondary care settings.

CONCLUSION

Pre‐tibial lacerations affect an increasingly ageing population, creating wounds that have poor healing rates. Initial management of these wounds has changed little in 20 years, despite significant improvements in wound management science and surgical techniques. This study identified 24 patients over a 3‐month period, one third of whom failed to heal over the study period. The data, limited by the sample size, identified that wound length was associated with non healing. Neither the comorbidities nor prescriptions explored showed any significant association although some seem to be more prevalent in this patient group. The study also highlighted other issues that require further exploration including the social and economic impact of these wounds on both patients and the health service at large.

ACKNOWLEDGEMENTS

We would like to thank all the staff of the two emergency departments who collected data and identified patients, and the community nursing teams for their enthusiasm in participating in this study. There were no competing interests in the development or the delivery of the study. There was no funding for this study.