ORIGINAL ARTICLE
In Search of the Optimal Processing Technique for Fat Grafting Rakesh Gupta, BA,* Matthew Brace, MD, MSc,† S. Mark Taylor, MD, FRCSC,† Michael Bezuhly, MD, MSc,‡§ and Paul Hong, MD, FRCSC†§ Objective: Unpredictability in graft retention remains a significant drawback of fat grafting. Processing of fat grafts has been the focus of several studies to improve graft survival. The objective of this study was to systematically review the outcomes of different fat graft processing techniques with the goal of (1) deriving clinically oriented insights and (2) identifying gaps in knowledge to stimulate future research. Methods: PubMed, EMBASE, and Cochrane Databases were searched to identify studies that compared different fat graft processing techniques. Outcome measures of interest were any subjective or objective measures of fat graft survival or reports of adverse events. Results: A total of 2056 abstracts were generated from the literature searches; 13 studies met the criteria for data extraction and analysis. Processing methods assessed included decantation, washing, gauze filtration, and centrifugation. Each processing method was found to be better than other methods, depending on the outcome measure used to study graft survival. As well, several studies found statistical equipoise in the outcome measures when analyzing the results of the different techniques. Adverse events were rarely reported and did not correlate with any processing method in particular. Conclusions: No firm concluding recommendation can be made to deem 1 processing technique superior to the others. However, it would seem that techniques, which use a combination of gentle washing and centrifugation, strike the optimal balance of preserving adipocyte viability while removing bulk of the contaminants. Key Words: Fat grafting, fat graft processing, fat graft survival, adipocyte survival (J Craniofac Surg 2015;26: 94–99)
F
at grafting is a common procedure with broad applications in craniofacial surgery. First described as a means of filling depressed forearm scars, the technique is now widely used, because in large part of its simplicity and low associated morbidity.1,2
From the *Faculty of Medicine and Divisions of †Otolaryngology-Head & Neck Surgery and ‡Plastic and Reconstructive Surgery, Department of Surgery, Dalhousie University; and §Dalhousie Pediatric Craniofacial Group, IWK Health Centre, Halifax, Nova Scotia, Canada. Received May 28, 2014. Accepted for publication July 31, 2014. Address correspondence and reprint requests to Paul Hong, MD, FRCSC, Dalhousie Pediatric Craniofacial Group, IWK Health Centre, 5850/5920 University Ave, PO Box 9700 Halifax, Nova Scotia, B3K 6R8 Canada; E-mail: [email protected] The authors report no conflicts of interest. Copyright © 2014 by Mutaz B. Habal, MD ISSN: 1049-2275 DOI: 10.1097/SCS.0000000000001259
94
Unfortunately, fat graft resorption remains a major limiting factor in the success of this procedure, to date, with studies reporting resorption rates of 20% to 90%.3 To minimize this drawback, attempts have been made to optimize each step of the transplantation procedure, including donor-site selection, harvesting, processing, and injection.4–6 The optimal method of processing remains one of the most controversial steps in the entire fat grafting procedure. Harvested fat grafts must be processed to separate the general components of the lipoaspirate—adipocytes, lipid, and serosangunious fluid—and obtain the highest concentration of intact, viable adipocytes for implantation.7 Several fat graft processing methods have been proposed to achieve this outcome including decantation, centrifugation, washing, filtration, and various combinations thereof. The purpose of this article is to systematically review original studies reporting on methods of fat graft processing. In doing so, it attempts to determine the optimal processing method for long-term graft retention to better inform the physician.
MATERIALS AND METHODS Protocol and Registration The statement of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses was followed in developing this review.8 A detailed research protocol was devised a priori to define the scope, objective, and methodology. The protocol was not registered or published.
Eligibility Criteria Studies were included if they met all of the following criteria: (1) assessment of different fat graft processing methods, (2) use of a comparison or control group, and (3) fat graft survival analysis. Studies were exlcuded if they met any of the following criteria: (1) testing of new or additional substances (including adiposederived stem cell enrichment techniques), (2) assessment of different settings or protocols within a single processing method, (3) assessment of different harvest or injection techniques, (4) assessment of different fat graft donor sites, and (5) retrospective reviews. Studies were required to meet criteria for evidence level 4 or higher, as defined by the Oxford Centre for Evidence-Based Medicine.9 These include case series, cohort studies, case-control studies, and randomized controlled trials. Animal model studies containing relevant data were also included. Only those articles published in English were searched; duplicate studies were excluded.
Information Sources A comprehensive literature search was performed using PubMed, EMBASE, and Cochrane Database of Systematic Reviews (Fig. 1). Titles and abstracts were obtained for all studies identified by the literature search. The bibliographies of those studies were hand-searched for any additional relevant studies not identified by the original database searches.
The Journal of Craniofacial Surgery • Volume 26, Number 1, January 2015
Copyright © 2014 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.
The Journal of Craniofacial Surgery • Volume 26, Number 1, January 2015
Processing Technique for Fat Grafting
Risk for Bias in Individual Studies Risk for selection and publication bias was evaluated for each study. Consideration was given to sample collection, inclusion and exclusion criteria, and outcome measures.
Summary Measures Measures were summarized as objective outcomes, subjective outcomes, and reports of adverse events. In cases where data were represented graphically rather than numerically, outcomes were summarized descriptively.
Synthesis of Results Data were analyzed according to the objective or subjective nature of the outcome measures. Meta-analysis was not performed because of the heterogeneity of specific outcomes across individual studies and the lack of access to raw data.
RESULTS Study Selection A total of 2056 articles were identified from the inital database and bibliography searches. A total of 18 articles met the criteria for full-text review. After full-text review, 5 additional articles were excluded for various reasons, which included (1) comparison of different settings/protocols within a single processing method, (2) comparison of preharvest graft handling, (3) a confounding centrifugation step in all groups, and (4) examination of harvested and processed fat without analysis of graft survival. Thirteen studies ultimately met inclusion criteria and were included in the final analysis. Figure 2 outlines the search and exclusion process.
Study Characteristics Characteristics of the 13 studies included in our analysis are shown in Table 1. Eleven studies were parallel group studies, and two were intrasubject (side versus side) comparison studies.
FIGURE 1. Details of search terms used in the systematic literature review.
Study Selection Three authors independently reviewed the list of titles and abstracts generated by the initial literature search and identified those that met the eligibility criteria. The other authors resolved any conflicting assessments. The full texts of these reports were then independently assessed for eligbility in a similar manner.
Data Extraction and Items The following data were extracted: study design, level of evidence, number of subjects, description of subjects, description of intervention, and outcomes reported. For quantitative outcome measures, mean or median values, SDs, confidence intervals, and P values were recorded when possible. Data obtained by comparable methodology were synthesized into tables as appropriate. When appropriate, quality assessment of included studies was performed using a 7-point quality assessment scoring system. One point was awarded for each of the following characteristics: (1) clearly stated inclusion/exclusion criteria; (2) validated assessment method; (3) appropriate comparison group (intrasubject or control group); (4) blinded observers; (5) adequate sample size or power analysis; (6) well-defined, consecutive samples; and (7) less than 5% lost to follow-up.
FIGURE 2. Flow diagram of article selection process.
© 2014 Mutaz B. Habal, MD
Copyright © 2014 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.
95
The Journal of Craniofacial Surgery • Volume 26, Number 1, January 2015
Gupta et al
TABLE 1. Characteristics of Studies Included in the Systematic Review Quality Score
n
4
25
3
14
N/A
10
Parallel group study
N/A
France
Parallel group study
Khater et al,19 2008
Bulgaria
Khater et al,20 2009
Study
Country
Study Design
Botti et al,17 2011
Italy, Romania
Butterwick,18 2002
United States
Condé-Green et al,13 2010
Brazil
Intrasubject side versus side comparison study Intrasubject side versus side comparison study Parallel group study
Condé-Green et al,14 2010
Brazil
Hoareau et al,15 2013
Fat Graft Characteristics Type: autograft; harvest site: abdomen, knee, thigh; recipient site: face Type: autograft; harvest site: medial knee, outer thigh, hip; recipient site: dorsal hand Type: not transplanted; harvest site: abdomen
20
N/A
Parallel group study
Bulgaria, France
Minn et al,11 2010
Subject Characteristics Age: mean of 46.3 y, range of 21–72 y; sex: 21 women
Processing Methods Examined
Outcomes
Filtration (strainer) and wash (0.9% of saline), centrifugation (3000 rpm for 3 min) Decantation, centrifugation (3600 rpm for 3 min)
Subjective: patient esthetic assessment
Age: range, 35–58 y; sex: 10 women
Decantation, centrifugation (3000 rpm for 3 min)
Type: not transplanted; harvest site: abdomen
Age: range, 28–64 y; sex: 20 women
36
Type: xenograft (human to mouse); harvest site: unknown; recipient site: dorsolateral region
Donor age: 43 (9) y; donor sex: 9 women; recipient: mice
3
30
Type: autograft; harvest site: subgluteal region; recipient site: face
Age: range, 15–47 y; sex: 26 women
Decantation, wash (0.9% of saline), centrifugation (3000 rpm for 3 min) Decantation, soft centrifugation (wash with RL, centrifuged 100 g for 1 s 2, then 400 g for 1 min), high-speed centrifugation (900 g for 3 min) Wash (saline), centrifugation (3000 rpm for 3 min)
Parallel group study
3
51
Type: autograft; harvest site: subgluteal region; recipient site: face
Age: mean of 32.68 y, range of 16–55 y; sex: 51 women
Wash (saline), centrifugation (3400 rpm for 3 min)
South Korea
Parallel group study
N/A
18
Donor age: unknown; donor sex: unknown; recipient: mice
Filtration (sieve), filtration (cotton gauze), centrifugation (1800 g for 3 min)
Ramon et al,12 200512
Israel
Parallel group study
N/A
22
Donor age: 32 y; donor sex: 1 woman; recipient: mice
Filtration (cotton towel), centrifugation (1500 rpm for 5 min 2)
Rohrich et al,3 2004
United States
Parallel group study
N/A
5
Age: unknown; sex: unknown
Decantation, centrifugation (500 g for 2 min)
Rose et al,7 2006
United States
Parallel group study
N/A
22
Type: xenograft (human to mouse); harvest site: transverse rectus abdominus muscle flaps; recipient site: nuchal area Type: xenograft (human to mouse); harvest site: breast; recipient site: nuchal subcutis Type: not transplanted; harvest site: abdomen, flank, thigh, medial knee Type: analyzed before transplant; harvest site: peri-umbilical region
Objective: aspirate intact adipocytes per HPF; subjective: aspirate gross appearance Objective: aspirate intact adipocytes per HPF Objective: oil and liquid released from adipose tissue after centrifugation, inflammatory markers; subjective: graft histologic appearance Subjective: physician and patient esthetic assessment; aspirate histologic appearance Objective: aspirate IHC of leptin and cyclin D1; subjective: physician, patient, and independent observer esthetic assessment, aspirate histologic appearance Objective: XTT assay absorbance, graft survival by weight; subjective: graft histologic appearance
Age: unknown; sex: unknown
Decantation, centrifugation (3000 rpm for 3 min), wash (0.9% of saline) and centrifugation (3000 rpm for 3 min)
Smith et al,4 2006
United States
Parallel group study
N/A
57
Type: xenograft (human to mouse); harvest site: abdomen; recipient site: dorsal flank skin
Donor age: unknown; donor sex: 3 women; recipient: mice
Zhu et al,16 201316
United States
Parallel group study
N/A
22
Type: not transplanted; harvest site: abdomen, hips, flanks
Age: mean of 45 y, range of 24–64 y; sex: 22 women
No preparation, centrifugation (500 g for 2 min), wash (RL), wash (0.9% of saline), wash (RL) and centrifugation (500 g for 2 min), wash (0.9% of saline) and centrifugation (500 g for 2 min) No preparation, decantation, centrifugation (3000 rpm for 3 min), wash (RL) and filtration
Age: mean of 53.5 y, range of 41–64 y; sex: 14 women
Subjective: physician and patient esthetic assessment
Objective: graft weight and volume; subjective: graft histologic appearance Objective: XTT assay absorbance
Objective: aspirate intact adipocyte per HPF, aspirate nucleated intact adipocyte per HPF, aspirate nucleated adipocyte cross-sectional area Objective: XTT assay absorbance, graft weight
Objective: aspirate percentage content of centrifugation density components; RBC, WBC, and growth factor content of grafts; relative lipolysis activity
IHC, immunohistochemistry; LR, lactated ringer; N/A, not applicable; XTT, (sodium 3’-[1-(phenylaminocarbonyl)-3,4-tetrazolium]-bis[4-methoxy-6-nitro] benzene sulfonic acid hydrate).
96
© 2014 Mutaz B. Habal, MD
Copyright © 2014 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.
The Journal of Craniofacial Surgery • Volume 26, Number 1, January 2015
Processing methods examined included decantation (including gravity separation and sedimentation), washing, gauze filtration (including concentration), and centrifugation. Four studies reported objective outcome measures, 3 studies reported subjective outcome measures, and 6 studies reported both. A total of 120 subjects were humans receiving autografts, and 133 were mice receiving xenografts from human donors. Five studies involved analysis of grafts without transplantation. A total of 177 donors were women and 8 donors were men, with an overall age range of 15 to 72 years. Donor characteristics were not reported in 3 studies. Donor sites included the breast, abdomen, flank, hip, knee, and thigh. Quality assessment scores were calculated in 4 studies; the remainder were considered experimental studies (Table 1).
Processing Technique for Fat Grafting
Objective Outcomes XTT Assay Absorbance
Overall Outcomes
Three studies assessed postprocessing adipocyte survival via XTT (sodium 3'-[1-(phenylaminocarbonyl)-3,4-tetrazolium]-bis[4methoxy-6-nitro] benzene sulfonic acid hydrate) absorbance assay, with increased absorbance indicating greater numbers of intact and metabolically active cells.10 Minn et al11 found that metal sieve concentrated samples exhibited significantly lower absorbance than those cotton gauze concentrated or centrifuged samples. Smith et al4 found that washed-plus-centrifuged samples showed significantly lower absorbance at 90 minutes than those washed alone, centrifuged alone, or unprocessed, with no significant differences among processed samples at 30, 60, or 120 minutes. Rohrich et al3 found no significant difference in absorption between decanted and centrifuged samples.
The summary of significant outcomes by processing methods is shown in Table 2.
Postgraft Weight
TABLE 2. Summary of the Significant Outcomes Processing Methods No preparation Decantation
Centrifugation
Wash
Filtration Wash and centrifugation Wash and filtration
Three studies assessed postgraft weight outcomes, with none finding signficant differences among any processing methods assessed.4,11,12
Outcomes Superior to multiple methods XTT assay absorbance4 Superior to centrifugation Aspirate intact adipocytes per HPF7,13 Aspirate nucleated adipocytes per HPF,7 serum IL-6 (high centrifugation only, P values not provided)15 Superior to wash Aspirate intact adipocytes per HPF14 Superior to wash and centrifugation Aspirate intact adipocytes per HPF7 Aspirate nucleated adipocytes per HPF7 Superior to multiple methods Aspirate percentage of aqueous liquid16 Superior to decantation Aspirate gross appearance of layer separation13 Graft histologic appearance of collagenous connective tissue integration15 Graft gross appearance of visible oil,15 serum MCP-1 (soft centrifugation only, P values not provided)15 Superior to wash Aspirate percentage of CD45+ cells (centrifugation pellet, P values not provided)14 Superior to wash and filtration Patient and observer esthetic assessment at 2 y17 Superior to centrifugation Aspirate intact adipocytes per HPF14 Aspirate percentage of CD45–CD31+ cells (P values not provided)14 Patient esthetic assessment at 1 y20 Independent observer esthetic assessment at 1 y20 Aspirate histologic appearance of collagenous connective tissue Preservation19,20 Superior to centrifugation Graft histologic appearance of fibrosis12 Inferior to multiple methods XTT assay absorbance at 90 min4 Superior to multiple methods Aspirate percentage of aqueous liquid16 Aspirate percentage of free lipid16 Aspirate relative to RBC concentration16 Aspirate relative to WBC concentration16 Aspirate agonist-induced lipolysis activity16
Intact Adipocytes per High-Powered Field Three studies assessed the number of intact adipocytes per high-powered field (HPF) on microscopy.7,13,14 Overall, the processing methods that involved less manipulation resulted in more intact adipocytes per HPF. Condé-Green et al13,14 found that decantation samples showed significantly more intact adipocytes per HPF than centrifuged or washed samples. Rose et al7 found that decantation samples demonstrated significantly more intact adipocytes per HPF than centrifuged or washed-plus-centrifuged samples.
Other Objective Outcomes Six studies assessed other objective outcomes. Condé-Green et al13,14 found that the pellet from centrifuged samples contained a higher fraction of hematopoietic (CD45+) and nonhematopoietic (CD45−) cells than the fat samples from the decantation group. This pellet had a high mesenchymal and endothelial cell content. The decantation group contained a higher fraction of CD45+ cells than the centrifuged samples with pellet removed. Washed samples contained a lower fraction of CD45+ and a higher fraction of nonhematopoietic CD45− cells than decanted samples.14 The authors concluded that use of the centrifuged pellet could enhance graft results because of its high content of mesenchymal stem cells but that washing produced the best sample in mesenchymal stem cell content free of hematopoietic contaminants. Hoareau et al15 found that high-speed centrifuged samples contained a 5-fold higher content of IL-6 relative to soft centrifuged samples, and both high-speed centrifuged and decanted samples contained twice the content of monocyte chemotactic protein-1 relative to soft centrifuged samples. These results suggested that transfer of fat grafts processed by high-speed centrifugation may lead to an increased host inflammatory response compared with that of decanted samples. Their results indicated that a “soft” centrifugation at 400 g combined with washing would optimize cell viability and purification of the grafts. Ramon et al12 found no significant difference in 16-week postgraft volume between gauze filtration and centrifuged samples. Rose et al7 showed that decantation-processed samples exhibited significantly more nucleated adipocytes per HPF than centrifuged samples, but no significant difference was seen in the adipocyte crosssectional area. Zhu et al16 found that washed-plus-filtered samples and centrifuged samples contained a signifiantly lower percentage of fluid content compared with control groups. They also found that washed-plus-filtered samples contained significantly lower percentage
© 2014 Mutaz B. Habal, MD
Copyright © 2014 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.
97
The Journal of Craniofacial Surgery • Volume 26, Number 1, January 2015
Gupta et al
of free lipid content and relative red blood cell and white blood cell concentrations than both centrifuged and control groups. Their conclusion was that washing and filtration produced the most viable and pure graft.16
Subjective Outcomes Esthetic Assessments Four studies assessed esthetic outcomes with varying results.17–20 Khater et al20 found that patients and independent observers preferred washed grafts to centrifuged grafts at 1-year follow-up but no significant differences were reported on physician assessments. In another study, no significant differences in patient or observer esthetic assessments were reported at 2-year follow-up across 10 specific facial regions, with the exception of the lips, for which the centrifugation group scored significantly higher than the filtration and wash group.17 The other 2 studies were purely descriptive with no quantitative or statistical analyses.18,19
Histologic and Gross Appearance Seven studies assessed histologic and gross appearance. Overall, the processing methods with less manipulation resulted in more intact adipocytes13,14 and greater integration with collagenous connective tissue.19,20 Minn et al11 found that metal sieve concentrated samples exhibited more inflammation than cotton gauze concentrated and centrifuged samples, with no differences in necrosis or vascularity. Ramon et al12 found that concentrated samples exhibited less fibrosis than centrifuged samples, with no differences in integrity, cyst/vaculoes, inflammation, or vascularity. CondéGreen et al14 found that centrifuged samples showed greater gross separation of layers than decanted samples. Hoareau et al15 found that decanted samples exhibited more grossly visible oil than centrifuged samples.
Adverse Events Five studies reported specific adverse events.4,12,18–20 Butterwick18 reported several transient adverse events in the decantation and centrifugation groups, including dusky gray-blue discoloration lasting 2 to 4 weeks (3/14 in both groups); dysesthesia lasting 2 to 3 hours (1/4 in the decantation group, 2/14 in the centrifugation group); and edema, discoloration, and surface irregularities lasting 3 to 6 weeks in both groups. They also reported infection in 1 of the 14 injections in the decantation group, which responded to a 10-day course of antibiotic therapy. Khater et al20 reported transient equivalent postoperative swelling between the centrifugation and wash groups. Two animals (one in the filtration group12, one in the centrifugation group4) in 2 different studies died before euthanization, but the cause was not explicitly attributed to fat grafting.
DISCUSSION Outcomes Analysis of fat graft outcomes with different processing techniques revealed heterogenous results because of the varying metric of assessments used in individual studies. Depending on the metric used, each processing method was found to be superior to the others in at least 1 study measure as outlined in Table 2. Many studies found statistical equipoise in the outcome measures when analyzing the results of the different techniques. As a result, no firm concluding recommendation can be made to deem 1 processing technique superior to the others. However, a number of themes are repeated in this review. Processing techniques that use the least amount of adipocyte manipulation results in a graft with the highest concentration of viable
98
cells, together with the highest amount of hematologic and liquid/ tumescence fluid contaminants. Whereas grafts subjected to higher centrifugation forces (>400 g) experience greater adipocyte loss and have the fewest number of intact cells, in the purest sample, that is free of contaminants. It would seem that techniques, which use a combination of gentle graft washing and centrifugation (
In Search of the Optimal Processing Technique for Fat Grafting Rakesh Gupta, BA,* Matthew Brace, MD, MSc,† S. Mark Taylor, MD, FRCSC,† Michael Bezuhly, MD, MSc,‡§ and Paul Hong, MD, FRCSC†§ Objective: Unpredictability in graft retention remains a significant drawback of fat grafting. Processing of fat grafts has been the focus of several studies to improve graft survival. The objective of this study was to systematically review the outcomes of different fat graft processing techniques with the goal of (1) deriving clinically oriented insights and (2) identifying gaps in knowledge to stimulate future research. Methods: PubMed, EMBASE, and Cochrane Databases were searched to identify studies that compared different fat graft processing techniques. Outcome measures of interest were any subjective or objective measures of fat graft survival or reports of adverse events. Results: A total of 2056 abstracts were generated from the literature searches; 13 studies met the criteria for data extraction and analysis. Processing methods assessed included decantation, washing, gauze filtration, and centrifugation. Each processing method was found to be better than other methods, depending on the outcome measure used to study graft survival. As well, several studies found statistical equipoise in the outcome measures when analyzing the results of the different techniques. Adverse events were rarely reported and did not correlate with any processing method in particular. Conclusions: No firm concluding recommendation can be made to deem 1 processing technique superior to the others. However, it would seem that techniques, which use a combination of gentle washing and centrifugation, strike the optimal balance of preserving adipocyte viability while removing bulk of the contaminants. Key Words: Fat grafting, fat graft processing, fat graft survival, adipocyte survival (J Craniofac Surg 2015;26: 94–99)
F
at grafting is a common procedure with broad applications in craniofacial surgery. First described as a means of filling depressed forearm scars, the technique is now widely used, because in large part of its simplicity and low associated morbidity.1,2
From the *Faculty of Medicine and Divisions of †Otolaryngology-Head & Neck Surgery and ‡Plastic and Reconstructive Surgery, Department of Surgery, Dalhousie University; and §Dalhousie Pediatric Craniofacial Group, IWK Health Centre, Halifax, Nova Scotia, Canada. Received May 28, 2014. Accepted for publication July 31, 2014. Address correspondence and reprint requests to Paul Hong, MD, FRCSC, Dalhousie Pediatric Craniofacial Group, IWK Health Centre, 5850/5920 University Ave, PO Box 9700 Halifax, Nova Scotia, B3K 6R8 Canada; E-mail: [email protected] The authors report no conflicts of interest. Copyright © 2014 by Mutaz B. Habal, MD ISSN: 1049-2275 DOI: 10.1097/SCS.0000000000001259
94
Unfortunately, fat graft resorption remains a major limiting factor in the success of this procedure, to date, with studies reporting resorption rates of 20% to 90%.3 To minimize this drawback, attempts have been made to optimize each step of the transplantation procedure, including donor-site selection, harvesting, processing, and injection.4–6 The optimal method of processing remains one of the most controversial steps in the entire fat grafting procedure. Harvested fat grafts must be processed to separate the general components of the lipoaspirate—adipocytes, lipid, and serosangunious fluid—and obtain the highest concentration of intact, viable adipocytes for implantation.7 Several fat graft processing methods have been proposed to achieve this outcome including decantation, centrifugation, washing, filtration, and various combinations thereof. The purpose of this article is to systematically review original studies reporting on methods of fat graft processing. In doing so, it attempts to determine the optimal processing method for long-term graft retention to better inform the physician.
MATERIALS AND METHODS Protocol and Registration The statement of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses was followed in developing this review.8 A detailed research protocol was devised a priori to define the scope, objective, and methodology. The protocol was not registered or published.
Eligibility Criteria Studies were included if they met all of the following criteria: (1) assessment of different fat graft processing methods, (2) use of a comparison or control group, and (3) fat graft survival analysis. Studies were exlcuded if they met any of the following criteria: (1) testing of new or additional substances (including adiposederived stem cell enrichment techniques), (2) assessment of different settings or protocols within a single processing method, (3) assessment of different harvest or injection techniques, (4) assessment of different fat graft donor sites, and (5) retrospective reviews. Studies were required to meet criteria for evidence level 4 or higher, as defined by the Oxford Centre for Evidence-Based Medicine.9 These include case series, cohort studies, case-control studies, and randomized controlled trials. Animal model studies containing relevant data were also included. Only those articles published in English were searched; duplicate studies were excluded.
Information Sources A comprehensive literature search was performed using PubMed, EMBASE, and Cochrane Database of Systematic Reviews (Fig. 1). Titles and abstracts were obtained for all studies identified by the literature search. The bibliographies of those studies were hand-searched for any additional relevant studies not identified by the original database searches.
The Journal of Craniofacial Surgery • Volume 26, Number 1, January 2015
Copyright © 2014 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.
The Journal of Craniofacial Surgery • Volume 26, Number 1, January 2015
Processing Technique for Fat Grafting
Risk for Bias in Individual Studies Risk for selection and publication bias was evaluated for each study. Consideration was given to sample collection, inclusion and exclusion criteria, and outcome measures.
Summary Measures Measures were summarized as objective outcomes, subjective outcomes, and reports of adverse events. In cases where data were represented graphically rather than numerically, outcomes were summarized descriptively.
Synthesis of Results Data were analyzed according to the objective or subjective nature of the outcome measures. Meta-analysis was not performed because of the heterogeneity of specific outcomes across individual studies and the lack of access to raw data.
RESULTS Study Selection A total of 2056 articles were identified from the inital database and bibliography searches. A total of 18 articles met the criteria for full-text review. After full-text review, 5 additional articles were excluded for various reasons, which included (1) comparison of different settings/protocols within a single processing method, (2) comparison of preharvest graft handling, (3) a confounding centrifugation step in all groups, and (4) examination of harvested and processed fat without analysis of graft survival. Thirteen studies ultimately met inclusion criteria and were included in the final analysis. Figure 2 outlines the search and exclusion process.
Study Characteristics Characteristics of the 13 studies included in our analysis are shown in Table 1. Eleven studies were parallel group studies, and two were intrasubject (side versus side) comparison studies.
FIGURE 1. Details of search terms used in the systematic literature review.
Study Selection Three authors independently reviewed the list of titles and abstracts generated by the initial literature search and identified those that met the eligibility criteria. The other authors resolved any conflicting assessments. The full texts of these reports were then independently assessed for eligbility in a similar manner.
Data Extraction and Items The following data were extracted: study design, level of evidence, number of subjects, description of subjects, description of intervention, and outcomes reported. For quantitative outcome measures, mean or median values, SDs, confidence intervals, and P values were recorded when possible. Data obtained by comparable methodology were synthesized into tables as appropriate. When appropriate, quality assessment of included studies was performed using a 7-point quality assessment scoring system. One point was awarded for each of the following characteristics: (1) clearly stated inclusion/exclusion criteria; (2) validated assessment method; (3) appropriate comparison group (intrasubject or control group); (4) blinded observers; (5) adequate sample size or power analysis; (6) well-defined, consecutive samples; and (7) less than 5% lost to follow-up.
FIGURE 2. Flow diagram of article selection process.
© 2014 Mutaz B. Habal, MD
Copyright © 2014 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.
95
The Journal of Craniofacial Surgery • Volume 26, Number 1, January 2015
Gupta et al
TABLE 1. Characteristics of Studies Included in the Systematic Review Quality Score
n
4
25
3
14
N/A
10
Parallel group study
N/A
France
Parallel group study
Khater et al,19 2008
Bulgaria
Khater et al,20 2009
Study
Country
Study Design
Botti et al,17 2011
Italy, Romania
Butterwick,18 2002
United States
Condé-Green et al,13 2010
Brazil
Intrasubject side versus side comparison study Intrasubject side versus side comparison study Parallel group study
Condé-Green et al,14 2010
Brazil
Hoareau et al,15 2013
Fat Graft Characteristics Type: autograft; harvest site: abdomen, knee, thigh; recipient site: face Type: autograft; harvest site: medial knee, outer thigh, hip; recipient site: dorsal hand Type: not transplanted; harvest site: abdomen
20
N/A
Parallel group study
Bulgaria, France
Minn et al,11 2010
Subject Characteristics Age: mean of 46.3 y, range of 21–72 y; sex: 21 women
Processing Methods Examined
Outcomes
Filtration (strainer) and wash (0.9% of saline), centrifugation (3000 rpm for 3 min) Decantation, centrifugation (3600 rpm for 3 min)
Subjective: patient esthetic assessment
Age: range, 35–58 y; sex: 10 women
Decantation, centrifugation (3000 rpm for 3 min)
Type: not transplanted; harvest site: abdomen
Age: range, 28–64 y; sex: 20 women
36
Type: xenograft (human to mouse); harvest site: unknown; recipient site: dorsolateral region
Donor age: 43 (9) y; donor sex: 9 women; recipient: mice
3
30
Type: autograft; harvest site: subgluteal region; recipient site: face
Age: range, 15–47 y; sex: 26 women
Decantation, wash (0.9% of saline), centrifugation (3000 rpm for 3 min) Decantation, soft centrifugation (wash with RL, centrifuged 100 g for 1 s 2, then 400 g for 1 min), high-speed centrifugation (900 g for 3 min) Wash (saline), centrifugation (3000 rpm for 3 min)
Parallel group study
3
51
Type: autograft; harvest site: subgluteal region; recipient site: face
Age: mean of 32.68 y, range of 16–55 y; sex: 51 women
Wash (saline), centrifugation (3400 rpm for 3 min)
South Korea
Parallel group study
N/A
18
Donor age: unknown; donor sex: unknown; recipient: mice
Filtration (sieve), filtration (cotton gauze), centrifugation (1800 g for 3 min)
Ramon et al,12 200512
Israel
Parallel group study
N/A
22
Donor age: 32 y; donor sex: 1 woman; recipient: mice
Filtration (cotton towel), centrifugation (1500 rpm for 5 min 2)
Rohrich et al,3 2004
United States
Parallel group study
N/A
5
Age: unknown; sex: unknown
Decantation, centrifugation (500 g for 2 min)
Rose et al,7 2006
United States
Parallel group study
N/A
22
Type: xenograft (human to mouse); harvest site: transverse rectus abdominus muscle flaps; recipient site: nuchal area Type: xenograft (human to mouse); harvest site: breast; recipient site: nuchal subcutis Type: not transplanted; harvest site: abdomen, flank, thigh, medial knee Type: analyzed before transplant; harvest site: peri-umbilical region
Objective: aspirate intact adipocytes per HPF; subjective: aspirate gross appearance Objective: aspirate intact adipocytes per HPF Objective: oil and liquid released from adipose tissue after centrifugation, inflammatory markers; subjective: graft histologic appearance Subjective: physician and patient esthetic assessment; aspirate histologic appearance Objective: aspirate IHC of leptin and cyclin D1; subjective: physician, patient, and independent observer esthetic assessment, aspirate histologic appearance Objective: XTT assay absorbance, graft survival by weight; subjective: graft histologic appearance
Age: unknown; sex: unknown
Decantation, centrifugation (3000 rpm for 3 min), wash (0.9% of saline) and centrifugation (3000 rpm for 3 min)
Smith et al,4 2006
United States
Parallel group study
N/A
57
Type: xenograft (human to mouse); harvest site: abdomen; recipient site: dorsal flank skin
Donor age: unknown; donor sex: 3 women; recipient: mice
Zhu et al,16 201316
United States
Parallel group study
N/A
22
Type: not transplanted; harvest site: abdomen, hips, flanks
Age: mean of 45 y, range of 24–64 y; sex: 22 women
No preparation, centrifugation (500 g for 2 min), wash (RL), wash (0.9% of saline), wash (RL) and centrifugation (500 g for 2 min), wash (0.9% of saline) and centrifugation (500 g for 2 min) No preparation, decantation, centrifugation (3000 rpm for 3 min), wash (RL) and filtration
Age: mean of 53.5 y, range of 41–64 y; sex: 14 women
Subjective: physician and patient esthetic assessment
Objective: graft weight and volume; subjective: graft histologic appearance Objective: XTT assay absorbance
Objective: aspirate intact adipocyte per HPF, aspirate nucleated intact adipocyte per HPF, aspirate nucleated adipocyte cross-sectional area Objective: XTT assay absorbance, graft weight
Objective: aspirate percentage content of centrifugation density components; RBC, WBC, and growth factor content of grafts; relative lipolysis activity
IHC, immunohistochemistry; LR, lactated ringer; N/A, not applicable; XTT, (sodium 3’-[1-(phenylaminocarbonyl)-3,4-tetrazolium]-bis[4-methoxy-6-nitro] benzene sulfonic acid hydrate).
96
© 2014 Mutaz B. Habal, MD
Copyright © 2014 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.
The Journal of Craniofacial Surgery • Volume 26, Number 1, January 2015
Processing methods examined included decantation (including gravity separation and sedimentation), washing, gauze filtration (including concentration), and centrifugation. Four studies reported objective outcome measures, 3 studies reported subjective outcome measures, and 6 studies reported both. A total of 120 subjects were humans receiving autografts, and 133 were mice receiving xenografts from human donors. Five studies involved analysis of grafts without transplantation. A total of 177 donors were women and 8 donors were men, with an overall age range of 15 to 72 years. Donor characteristics were not reported in 3 studies. Donor sites included the breast, abdomen, flank, hip, knee, and thigh. Quality assessment scores were calculated in 4 studies; the remainder were considered experimental studies (Table 1).
Processing Technique for Fat Grafting
Objective Outcomes XTT Assay Absorbance
Overall Outcomes
Three studies assessed postprocessing adipocyte survival via XTT (sodium 3'-[1-(phenylaminocarbonyl)-3,4-tetrazolium]-bis[4methoxy-6-nitro] benzene sulfonic acid hydrate) absorbance assay, with increased absorbance indicating greater numbers of intact and metabolically active cells.10 Minn et al11 found that metal sieve concentrated samples exhibited significantly lower absorbance than those cotton gauze concentrated or centrifuged samples. Smith et al4 found that washed-plus-centrifuged samples showed significantly lower absorbance at 90 minutes than those washed alone, centrifuged alone, or unprocessed, with no significant differences among processed samples at 30, 60, or 120 minutes. Rohrich et al3 found no significant difference in absorption between decanted and centrifuged samples.
The summary of significant outcomes by processing methods is shown in Table 2.
Postgraft Weight
TABLE 2. Summary of the Significant Outcomes Processing Methods No preparation Decantation
Centrifugation
Wash
Filtration Wash and centrifugation Wash and filtration
Three studies assessed postgraft weight outcomes, with none finding signficant differences among any processing methods assessed.4,11,12
Outcomes Superior to multiple methods XTT assay absorbance4 Superior to centrifugation Aspirate intact adipocytes per HPF7,13 Aspirate nucleated adipocytes per HPF,7 serum IL-6 (high centrifugation only, P values not provided)15 Superior to wash Aspirate intact adipocytes per HPF14 Superior to wash and centrifugation Aspirate intact adipocytes per HPF7 Aspirate nucleated adipocytes per HPF7 Superior to multiple methods Aspirate percentage of aqueous liquid16 Superior to decantation Aspirate gross appearance of layer separation13 Graft histologic appearance of collagenous connective tissue integration15 Graft gross appearance of visible oil,15 serum MCP-1 (soft centrifugation only, P values not provided)15 Superior to wash Aspirate percentage of CD45+ cells (centrifugation pellet, P values not provided)14 Superior to wash and filtration Patient and observer esthetic assessment at 2 y17 Superior to centrifugation Aspirate intact adipocytes per HPF14 Aspirate percentage of CD45–CD31+ cells (P values not provided)14 Patient esthetic assessment at 1 y20 Independent observer esthetic assessment at 1 y20 Aspirate histologic appearance of collagenous connective tissue Preservation19,20 Superior to centrifugation Graft histologic appearance of fibrosis12 Inferior to multiple methods XTT assay absorbance at 90 min4 Superior to multiple methods Aspirate percentage of aqueous liquid16 Aspirate percentage of free lipid16 Aspirate relative to RBC concentration16 Aspirate relative to WBC concentration16 Aspirate agonist-induced lipolysis activity16
Intact Adipocytes per High-Powered Field Three studies assessed the number of intact adipocytes per high-powered field (HPF) on microscopy.7,13,14 Overall, the processing methods that involved less manipulation resulted in more intact adipocytes per HPF. Condé-Green et al13,14 found that decantation samples showed significantly more intact adipocytes per HPF than centrifuged or washed samples. Rose et al7 found that decantation samples demonstrated significantly more intact adipocytes per HPF than centrifuged or washed-plus-centrifuged samples.
Other Objective Outcomes Six studies assessed other objective outcomes. Condé-Green et al13,14 found that the pellet from centrifuged samples contained a higher fraction of hematopoietic (CD45+) and nonhematopoietic (CD45−) cells than the fat samples from the decantation group. This pellet had a high mesenchymal and endothelial cell content. The decantation group contained a higher fraction of CD45+ cells than the centrifuged samples with pellet removed. Washed samples contained a lower fraction of CD45+ and a higher fraction of nonhematopoietic CD45− cells than decanted samples.14 The authors concluded that use of the centrifuged pellet could enhance graft results because of its high content of mesenchymal stem cells but that washing produced the best sample in mesenchymal stem cell content free of hematopoietic contaminants. Hoareau et al15 found that high-speed centrifuged samples contained a 5-fold higher content of IL-6 relative to soft centrifuged samples, and both high-speed centrifuged and decanted samples contained twice the content of monocyte chemotactic protein-1 relative to soft centrifuged samples. These results suggested that transfer of fat grafts processed by high-speed centrifugation may lead to an increased host inflammatory response compared with that of decanted samples. Their results indicated that a “soft” centrifugation at 400 g combined with washing would optimize cell viability and purification of the grafts. Ramon et al12 found no significant difference in 16-week postgraft volume between gauze filtration and centrifuged samples. Rose et al7 showed that decantation-processed samples exhibited significantly more nucleated adipocytes per HPF than centrifuged samples, but no significant difference was seen in the adipocyte crosssectional area. Zhu et al16 found that washed-plus-filtered samples and centrifuged samples contained a signifiantly lower percentage of fluid content compared with control groups. They also found that washed-plus-filtered samples contained significantly lower percentage
© 2014 Mutaz B. Habal, MD
Copyright © 2014 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.
97
The Journal of Craniofacial Surgery • Volume 26, Number 1, January 2015
Gupta et al
of free lipid content and relative red blood cell and white blood cell concentrations than both centrifuged and control groups. Their conclusion was that washing and filtration produced the most viable and pure graft.16
Subjective Outcomes Esthetic Assessments Four studies assessed esthetic outcomes with varying results.17–20 Khater et al20 found that patients and independent observers preferred washed grafts to centrifuged grafts at 1-year follow-up but no significant differences were reported on physician assessments. In another study, no significant differences in patient or observer esthetic assessments were reported at 2-year follow-up across 10 specific facial regions, with the exception of the lips, for which the centrifugation group scored significantly higher than the filtration and wash group.17 The other 2 studies were purely descriptive with no quantitative or statistical analyses.18,19
Histologic and Gross Appearance Seven studies assessed histologic and gross appearance. Overall, the processing methods with less manipulation resulted in more intact adipocytes13,14 and greater integration with collagenous connective tissue.19,20 Minn et al11 found that metal sieve concentrated samples exhibited more inflammation than cotton gauze concentrated and centrifuged samples, with no differences in necrosis or vascularity. Ramon et al12 found that concentrated samples exhibited less fibrosis than centrifuged samples, with no differences in integrity, cyst/vaculoes, inflammation, or vascularity. CondéGreen et al14 found that centrifuged samples showed greater gross separation of layers than decanted samples. Hoareau et al15 found that decanted samples exhibited more grossly visible oil than centrifuged samples.
Adverse Events Five studies reported specific adverse events.4,12,18–20 Butterwick18 reported several transient adverse events in the decantation and centrifugation groups, including dusky gray-blue discoloration lasting 2 to 4 weeks (3/14 in both groups); dysesthesia lasting 2 to 3 hours (1/4 in the decantation group, 2/14 in the centrifugation group); and edema, discoloration, and surface irregularities lasting 3 to 6 weeks in both groups. They also reported infection in 1 of the 14 injections in the decantation group, which responded to a 10-day course of antibiotic therapy. Khater et al20 reported transient equivalent postoperative swelling between the centrifugation and wash groups. Two animals (one in the filtration group12, one in the centrifugation group4) in 2 different studies died before euthanization, but the cause was not explicitly attributed to fat grafting.
DISCUSSION Outcomes Analysis of fat graft outcomes with different processing techniques revealed heterogenous results because of the varying metric of assessments used in individual studies. Depending on the metric used, each processing method was found to be superior to the others in at least 1 study measure as outlined in Table 2. Many studies found statistical equipoise in the outcome measures when analyzing the results of the different techniques. As a result, no firm concluding recommendation can be made to deem 1 processing technique superior to the others. However, a number of themes are repeated in this review. Processing techniques that use the least amount of adipocyte manipulation results in a graft with the highest concentration of viable
98
cells, together with the highest amount of hematologic and liquid/ tumescence fluid contaminants. Whereas grafts subjected to higher centrifugation forces (>400 g) experience greater adipocyte loss and have the fewest number of intact cells, in the purest sample, that is free of contaminants. It would seem that techniques, which use a combination of gentle graft washing and centrifugation (
