The effectiveness of pressure therapy (15–25 mmHg) for hypertrophic burn scars: A systematic review and meta-analysis

Although pressure therapy (PT) represents the standard care for prevention and treatment of hypertrophic scar (HS) from burns, its practice is largely based on empirical evidence and its effectiveness remains controversial. To clarify the effect of PT (15–25 mmHg) for HS, we performed the systematic review and meta-analysis. Several electronic databases were screened to identify related randomized controlled trials (RCTs). 12 RCTs involving 710 patients with 761 HS resulting from burn injuries were included. Compared with non/low-PT, cases treated with PT (15–25 mmHg) showed significant differences in Vancouver Scar Scale score (MD = −0.58, 95% CI = −0.78–−0.37), thickness (SMD = −0.25, 95% CI = −0.40–−0.11), brightness (MD = 2.00, 95% CI = 0.59–3.42), redness (MD = −0.79, 95% CI = −1.52–−0.07), pigmentation (MD = −0.16, 95% CI = −0.32–−0.00) and hardness (SMD = −0.65, 95% CI = −1.07–−0.23). However, there was no difference in vascularity (MD = 0.03, 95% CI = −0.43–0.48). Our analysis indicated that patients with HS who were managed with PT (15–25 mmHg) showed significant improvements. Due to limitations, more large and well-designed studies are needed to confirm our findings and the side-effects of the PT may also need to be evaluated.

Data extraction. Two investigators independently extracted the following items from each eligible studies: (1) basic information: Surname of the first author, country of the investigation and year of publication; (2) characteristics of participants in each study group, including sample size, average age, ethnicity, percentage of total body surface area (%TBSA) and burn site; (3) treatment information: type of interventions, pressure dose, wear time, duration of treatment, follow-up period, number of cases that discontinued treatment and those that were lost during follow-up; (4) clinical outcomes. In case of any conflict, a discussion was carried out to achieve consensus.
Methodological quality assessment. The methodological quality of each included study was evaluated using the Cochrane Risk of Bias tool 26 . The items included random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias) and other bias. Two authors assessed the quality of eligible studies independently and discrepancies were resolved by discussion. Data analysis. Statistical analysis was undertaken using RevMan5.3 software (Cochrane Collaboration, Copenhagen, The Nordic Cochrane Centre). Inverse variance method in random-effects model was applied in the meta-analysis. Heterogeneity was estimated using the I 2 statistic and Q tests, I 2 < 50% and P > 0.1 indicated low risk of heterogeneity. For each of the comparisons, effect sizes for studies using the same outcome measures were pooled using weighted mean difference (MD) and where not applicable, standardized mean difference (SMD) was used. 95% confidence interval (CI) was also calculated for each pooled result. The Z test was used to assess the statistical significance of the pooled MD/SMD and two-tailed P < 0.05 was considered significant. Sensitivity analyses were performed by using fixed-effect model to evaluate the stability of the result. We constructed funnel plot to assess for the possibility of publication bias. The resulting symmetrical funnel plot indicated low risk of publication bias.

Results
Study selection. Our search results yielded 623 records. After excluding duplications, we screened 457 titles and abstracts. A total of 420 records were deleted. This included 258 irrelevant studies, 29 meta-analyses, systematic reviews and reviews, 69 animal studies, 64 editorials and case reports. 37 potentially relevant studies were left for further review of the full-text. We continued to exclude 25 articles: 15 studies for non-RCT study design, 4 studies for insufficient information, 2 brief reports and 4 duplicated reports. 12 studies 27-38 were ultimately included for meta-analysis. The details of study selection is shown in Fig. 1 31 . Two studies 27,30 used within-patient design where each patient had at least two wounds, some wounds were/one wound was treated with PT (15-25 mmHg), others wounds/one wound as a control; or in patients with only one wound, one-half of the wound applied PT (15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25) and the other-half was considered as control. Ten studies 28,29,[31][32][33][34][35][36][37][38] used between-patient design where each patient was given a treatment. 441, 171, 226 HS were allocated into those treated with PT (15-25 mmHg), low-PT and non-PT groups respectively. Seven of these studies 27,31,32,[35][36][37][38] began PT for burn patients 1 to 3 weeks after wound closure or reepithelialisation. Two studies 28,29 started PT for patients several months post-injury. And three trials 33,34,36 did not report the time interval between PT and time of injury. Eleven [27][28][29][31][32][33][34][35][36][37][38] of the twelve trials reported the age of patients, two trials 33,34 included only pediatric patients, five 27,29,31,32,37 studied adults (age ≥ 18 years) and four 28,35,36,38 investigated both children and adults. Two studies 34,36 did not mention the burn site while the others were on the limbs. Four studies [27][28][29]36 reported burn surface area as size dimension, six [31][32][33][34][35]37 used %TBSA, while it was not reported in two studies 30,38 . Different kinds of pressure garments such as: Medical Z Corporation ® , Jobst ® , Urgosyval ® , Tricolast ® , etc were used in those studies. Four studies 31,36-38 did not clearly describe the manufacturer of pressure garment. Only five studies 27,29,30,32,33 reported the pressure value (scar/garment interface) assessment method. The longest duration of treatment was 12 months 30 and the shortest 2 months 36 , while one study 31 did not report the duration of treatment. Endpoint outcomes included VSS score, thickness, color, pigmentation, hardness and vascularity. Three studies 27,28,35 had cases lost in the follow-up period and one study 28 inadequately reported the reasons for losing. The risk of bias figure and the risk of bias summery figure are shown in Fig. 2A,B. Overall, the quality of included studies was under moderate risk of bias.

Meta-analysis.
The results of meta-analysis are displayed in Table 2. VSS score. VSS score consists of four parameters-pliability, height, vascularity and pigmentation. It is a widely used tool to assess the severity of scar 36 . The total score is 15 and normal skin scores 0. The higher the score, the more severe the scar. Five studies 33,34,36,38 involving 260 participants reported VSS score. The combined MD showed that PT (15-25 mmHg) could significantly reduce the VSS score (I 2 = 37%, MD = − 0.60, 95% CI = − 0.92-− 0.28, P < 0.01, random-effect model). The result varied little when a fixed-effect model was applied (MD = − 0.58, 95% CI = − 0.78-− 0.37, P < 0.01) (Fig. 3), which suggested that the result was stable.
Other outcomes. One of the studies 31 with 122 burn patients (64 in PT group and 58 in non-PT) that was undertaken to determine the efficacy of PT (15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25) showed no significant differences between the two groups when the length of hospital stay and time for wound maturation were compared. Another study 28 concerned with the scar pain and itch (using Visual Analog Scale, VAS) found that PT had more effectiveness in reducing scar pain than non-PT (P = 0.02). However, there was no significant difference in alleviating itching between the two groups 28 .   Additional analysis. An unpublished trial was included in previous meta-analysis 21 , but we did not obtain the trial in all of our available databases. So, we performed an additional analysis, extracting the trial's data from the previous study 21 and recalculating the results. As shown in Supplementary File S2, all related clinical outcomes were unchanged.
Publication bias. Publication bias was assessed by funnel plot. All funnel plots were generally symmetrical (See Supplementary Fig. S1), indicating low risk of publication bias in our study.

Discussion
PT represents the standard care for prevention and treatment of HS from burns, but there is no scientific evidence for using it 22,23 . Practice is largely based on empirical evidence and its effectiveness remains controversial in current studies 22 . This makes it unfavorable for clinicians to choose effective treatment measures. In this systematic review and meta-analysis, we included all relevant RCTs to compare the therapeutic effects of PT (15-25 mmHg) and non/low-PT. The results indicated that PT (15-25 mmHg) could improve clinical effects including decreasing VSS score, pigmentation, redness and increasing scar brightness. Our results are reliable given that pooled results were unaltered in sensitivity and additional analyses and there was no publication bias in the included studies. However, for increasing scar yellowness, the result was not stable enough in sensitivity analysis and needs to be investigated further. Similarly, the results of scar thickness, hardness and vascularity may be influenced by the study with no data or inconsistent outcome indicators. One study 30 that provided no data gave results on scar thickness and hardness in patients who were treated with PT and non-PT for 12 months. The study indicated that PT had positive therapeutical effect at 9 months post-burn, but showed no difference between PT and non-PT at 12 months. The combined result found no significant difference between PT and non/low -PT in vascularity (MD = 0.03, 95% CI = − 0.43-0.48, P = 0.91). One study 35 measured the changes of scar tissue blood perfusion by laser Doppler perfusion imaging and it showed that PT (15-25 mmHg) could significantly reduce the blood perfusion of scar tissue and thus reduce the scar vascularity 35 . However, this study data cannot be pooled.
The overall methodological quality of the included studies was moderate. Eight of the twelve studies did not report the method of randomization and ten had inadequate reports of allocation concealment. There was inadequate blinding in most studies and as previous researches have demonstrated inadequate blinding maybe associated with exaggerated treatment effects 21 . One study 28 inadequately reported the reasons for withdrawal from treatment. It has been previously shown that the treatment efficacy in patients who withdrew from researches   were different from those that completed the trial. Studies that failed to provide details regarding withdrawals were at an increasing risk of producing invalid results 26 . Future investigations should ensure adequate randomization, concealment of allocation, blinding of patients and outcome assessors and descriptions of withdrawals and losing.
To be effective, pressure garment should be worn for at least 23 hours a day and the treatment should be continued for a period of 6-12 months or until the scar matures 9,22,[39][40][41] . However, the exact optimal pressure required for effective treatment has never been scientifically established 15 . 15-25 mmHg pressure is used mostly in clinical practice. Higher pressures increase the effect and it has also been claimed to give more rapid results in the time of scar maturation 19 . However, pressure above 40 mmHg induces discomfort and potential harm, such as blistering, paresthesia, abnormal bone growth, limb necrosis, etc 12,42 . Moreover, higher pressure may also cause higher risk of pressure loss over time and increase the incidence of incompliance 6,43 . On the contrary, pressure less than 15 mmHg may appear to have poor/no effect 2,28,29 . A 12-year long-term high-quality study 27 demonstrated that 15mmHg was considered to be the minimum effective "dose". And other previous studies 29,32 suggested that the pressure level required might need to be higher than 15 mmHg. Generally, it is recommended that pressure should be maintained between 15 to 25 mmHg which being above capillary pressure, diminishes the supply of blood and nutrients to the scar tissue and is safer and more effective 2,27 . Thus, in this study, we took the non-PT and low-PT as one treatment group to assess the effect of PT (15-25 mmHg) for hypertrophic burn scars.
Although we found that PT (15-25 mmHg) had positive impacts on HS, some confounding factors could influence the therapeutic effects. Firstly, the time to begin PT 39,40,[44][45][46] . Different time interval between PT and burn-injury may have different PT effects, for example a recently healed burn wound and an established HS may respond differently to PT (15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25). Previous studies recommended that PT should be used as soon as the healing skin can tolerate the pressure 39,40,44 and/or shear force generated by the intervention 45,46 . The earlier the treatment begins the better the outcomes 44 . Secondly, the age of the burn patient [47][48][49] . Different life stages may have different healing processes and treatment compliances, which may have a great influence on assessing the effect of PT 48 . Thirdly, the location of burn 12,50 . Burn in different parts of the body may need different treatment requirements. And pressure is difficult to apply evenly across the body, particularly in concave areas and flexor joints 12 . Fourthly, methods of measuring the outcomes. Inequable validity may be found in different evaluation methods or scales 36 . Last but not least, pressure monitoring. In the course of PT, pressure losing along with therapy could also impact the final treatment effect and thus more attention should be paid on pressure monitoring 29,32 . However, in our included studies, the confounding factors were different or unreported and due to insufficient information we could not consider their influence. The effect of PT at different terms post-injury, life stage of patients, parts of the body burned, depth of burn, etc., should be further investigated in future studies. In addition, recognized and/or objective evaluation methods should be utilized and the pressure in scar/garment interface should be monitored in the future investigations.
Several investigations 48,51,52 showed that PT had effectiveness in enhancing HS maturation and controlling the itch associated with HS because PT could reduce collagen synthesis and prevent HS formation and contracture. However, others studies 49,53 found PT may cause skin breakdown/ulceration and aggravate the level of itching, especially in summer months. Prolonged wear of pressure garment may affect the skin perspiration and thus create heat and pruritus 49 . Only one of the included studies 31 focused on length of hospital stay and time of wound maturation for which no significant differences were found between the PT and non-PT group. Another study 28 found that PT had no more effectiveness in alleviating scar itching than non-PT where VAS score of itching changed from 4.47 ± 2.45 to 2.63 ± 1.91 and from 4.78 ± 3.35 to 3.09 ± 2.34 in PT and non-PT group respectively. Due to limited number of studies the efficacy of PT in scar maturation and itching could not be assessed to the best of its potential.
Some side-effects of PT such as blistering, ulceration, scar breakdown, limb swelling, etc. which are usually caused by too much pressure being applied were reported in previous studies 12,54,55 . In all of our included studies, 15-25 mmHg pressure, which was considered as safety pressure level was used in PT group. However, other adverse effects such as skeletal deformity, heterotopic ossification, muscle atrophy, joint stiffness 47,56 , etc. that could be caused by long-term usage of pressure garment may not have occurred or were unreported in our included studies. Thus, we did not assess the safety of PT. Future studies may need to pay more attention on the adverse-effects of PT. In addition, some studies 12,53 declared that poor appearance of pressure garments made patients feel self-conscious, causing embarrassment and other problems to patients. More RCTs may also need to investigate the impact of PT in patients' quality of life. In this way, the effects of PT may be investigated more comprehensively and accurately.
The exact mechanisms of how pressure positively influences the scar outcome following burns are not fully understood [7][8][9] . However, it is widely believed that the pressure can control collagen synthesis, facilitate scar maturation and reduce scar redness by limiting the supply of blood, oxygen and nutrients to the scar tissue 28,29 . Pressure has been also postulated to reduce the levels of collagen production more rapidly than the natural maturation process 15 . Mechanical loading induces alteration in collagen fiber turnover, remodeling and realignment and the reduced development of whorled collagen nodules result in the thinning and softening of scar tissues 7,12,16 . In addition, it is accepted that application of pressure commonly alleviates the itch and pain associated with active hypertrophic scars 42,43 .
In 2009, a standard meta-analysis 21 , included six RCTs, involving 316 patients, pooling four studies' data suggested that PT when compared non/low-PT could only mildly improve scar height (SMD = − 0.31, 95% CI = − 0.63 to 0.00, P = 0.05), but could not to alter global scar scores (VSS), pigmentation, vascularity, pliability and colour. Based on the current inconsistent study conclusions and the results of Meta-analysis, an international clinical recommendation 23 in 2014 concluded: "The long-time standard care for prevention and treatment of hypertrophic scars from burns is largely based on empirical evidence; no change in global scar scores and only small improvement in scar height was reported in meta-analysis; low pressure is less effective than high-pressure treatments; and patients with moderate or severe scarring experienced greater clinical benefit" 23 . And another useful guide 22 commented: "PT alleviates itching and pain associated with abnormal scars but with no scientific evidence in the use of pressure garments; given current lack of evidence, well-designed clinical trials are required to examine the effectiveness, risks and costs of PT" 22 .
The sample-size of previous meta-analysis 21 was relatively small and the conclusions were not stable enough. From 2009 to now, five new relevant RCTs were carried out. Therefore, we reappraised the effect of PT (15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25) for HS in burns. The results of our study were significantly different from the previous meta-analysis. The sample-size was larger, the pooled results were unaltered in sensitivity analysis and there was no obvious publication bias in the included studies. More interestingly, we added an unpublished trial data in our study and all related clinical outcomes were found to be unchanged. So, our results are more reliable.
However, it is important to note the limitations of our study. Firstly, our results may be influenced by the small number of included studies, the limited sample-size and inconsistent clinical outcomes of each study. Secondly, due to insufficient data, our study did not consider the %TBSA, burn degree and burn site although the different %TBSA, burn degree and burn site may have varying efficacy. Thirdly, since long-term follow-up studies were rare, our study failed to analyze the prospective efficacy of PT. Fourthly, none of the included studies studied adverse effects and we were unable to assess the safety of PT. Last but not least, our analysis suffered in quality of included studies because most of the studies did not describe the allocation concealment and blinding method, which may exaggerate the treatment effects, especially in subjective outcomes.

Conclusions
Our meta-analysis demonstrated that burn patients managed with PT (15-25 mmHg) showed significant improvements. Due to the limitations of the current studies, larger and well-designed studies are needed to confirm our findings. Furthermore, the side-effects of PT may also need to be evaluated in the future.