Proportionate mortality following dysvascular partial foot amputation and how this compares to transtibial amputation: a systematic review

Abstract

Purpose

A large proportion of people die in the years following dysvascular partial foot amputation (PFA) or transtibial amputation (TTA) given the long-term consequences of peripheral vascular disease and/or diabetes. A critical appraisal of recent research is needed to understand the underlying cause of variation and synthesise data for use in consultations about amputation surgery and patient-facing resources. This systematic review aimed to describe proportionate mortality following dysvascular PFA and to compare this between PFA and TTA.

Materials and methods

The review protocol was registered in PROSPERO (CRD42023399161). Peer-reviewed studies of original research were included if they: were published in English between 1 January 2016, and 12 April 2024, included discrete cohorts with PFA, or PFA and TTA, and measured proportionate mortality following dysvascular amputation.

Results

Seventeen studies were included in the review. Following dysvascular PFA, proportionate mortality increased from 30 days (2.1%) to 1-year (13.9%), 3-years (30.1%), and 5-years (42.2%). One study compared proportionate mortality 1-year after dysvascular PFA and TTA, showing a higher relative risk of dying after TTA (RR 1.51).

Conclusions

Proportionate mortality has not changed in recent years. These results are comparable to a previous systematic review that included studies published before 31 December 2015.

Implications for rehabilitation

• It is important to ensure data describing mortality in the years following dysvascular partial foot or transtibial amputation is up to date and accurate.

• Evidence about proportionate mortality has not changed in recent years and the results are comparable to previous systematic reviews.

• Data describing mortality outcomes can be used in decision aids that support conversations about the choice of amputation level.

Keywords:

• Amputation
• partial foot
• transtibial
• peripheral vascular disease
• mortality

Introduction

Dysvascular partial foot amputation (PFA) or transtibial amputation (TTA) is a common sequel to the long-term complications of peripheral vascular disease (PVD) and/or diabetes that leads to a reduction of blood supply to a body part and complications such as ulceration, skin breakdown, and infection [1–3].

Many people who undergo dysvascular PFA or TTA are older and have multiple comorbidities such as renal disease, or coronary heart disease [1,4]. As such, a large proportion of people die in the years following dysvascular PFA or TTA [3]. For example, based on a systematic review that included studies published until the 31 December 2015 [5], the proportion of people that died following dysvascular PFA increased from 30 days (2.8%, 95%CI 2.2–3.6) to 1-year (17.3%, 95%CI 14.6–20.2), 3-years (30.6%, 95%CI 23.0–38.7), and 5-years (41.2%, 95%CI 32.6–50.1). In comparison, those people who underwent TTA had a significantly greater relative risk (RR) of dying at 30 days (RR 2.6, 95%CI 1.6–4.1, p < 0.001), 1-year (RR 1.5, 95%CI 1.4–1.6, p < 0.001), and 5-years (RR 1.3, 95%CI 1.2–1.5, p < 0.001); noting there were no data comparing the RR of dying 3-years after PFA or TTA [5]. The greater RR of dying following TTA was thought to reflect that people considered suitable candidates for TTA likely had more advanced systemic disease which made PFA unsuitable [6–8].

In recent years, some studies have reported that the proportion of people who died in the years following dysvascular PFA and TTA has declined [9–11]. By contrast, others suggest that proportionate mortality has increased [7,12–16]. To understand the variation between these studies, a critical appraisal is necessary to understand how different methodologies and populations may have influenced these results. Only then can we determine whether proportionate mortality following dysvascular PFA has changed in recent years and how the RR compares to TTA.

Accurate and up-to-date evidence about proportionate mortality following dysvascular PFA, and how this compares to TTA, is important given these data are used to counsel people about the likely outcomes of these different surgeries, and thereby help inform difficult decisions about the choice of amputation level [5,17–27]. These data are also incorporated into patient-facing resources, such as decision aids, that present unbiased information about the treatment options (e.g., PFA vs. TTA) and the likely outcomes of each (e.g., proportionate mortality over time) [5,17–27]. Given this background, the aims of this systematic review were to:

1. Describe proportionate mortality following dysvascular PFA.

2. Compare proportionate mortality between dysvascular PFA and TTA.

Methods

This systematic review used the same methods as reported in the published protocol [28] and related peer reviewed research [5] and as such, a summary of the methods is reported here. This systematic review was registered with PROSPERO (CRD42023399161) and is reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [29] including a copy of the PRISMA checklist (Additional File 1).

Search strategy

A systematic search of the literature was conducted using MEDLINE, EMBASE, PsycINFO, AMED, CINAHL, ProQuest Nursing and Allied Health, and Web of Science. A list of search terms related to the amputation level and mortality, along with their synonyms and acronyms, were used in conjunction with wildcards and Boolean operators as part of a title, abstract, and keyword search. As the restriction of non-English studies does not seem to alter the outcome of systematic reviews and meta-analyses, the search was limited to the English language [30,31]. Searches were limited to studies published between 1 January 2016 and 12 April 2024, given a preceding systematic review included studies published to 31 December 2015 [5]. In keeping with the PRISMA guidelines [29], an illustrative search is presented for one database (Table 1) with the remaining searches included in Additional File 2. The search strategy was rigorously developed and tested as detailed in the systematic review protocol [28].

Table 1. Example search for CINAHL database.

Search	Field code	Search term(s)
1	MH	“Amputation”
2	MH	“Amputees”
3	TI, AB, SU	Amput* AND (Major OR Lowerlimb* OR “Lower limb*” OR “Lower extremit*” OR “Limb loss” OR LEA OR LLA)
4	TI, AB, SU	Amput* AND (Transtibial OR “Trans tibial” OR Belowknee OR Below W2 knee OR “below knee” OR TTA OR BKA)
5	TI, AB, SU	Amput* AND (Minor OR "partial foot" OR Chopart* OR Lisfranc* OR Tarsometatarsal OR Transmetatarsal OR Midtarsal OR "mid tarsal" OR Midfoot OR "mid foot" OR Ray OR Phalangeal OR Metatarsophalangeal OR Toe* OR Transtarsal OR "trans tarsal" OR TMT OR TMA OR MTP OR PFA)
6		1 OR 2 OR 3 OR 4 OR 5
7	TI, AB, SU	(Mortality OR Death OR Survival) AND (Rate* OR Proport* OR Percent*)
8		6 AND 7
9		Limit 8 to English Language
10		Limit 9 to publication date: 1 January 2016 to 12 April 2024
11		Limit 10 to peer-reviewed, academic journals

Field codes: MH exact major and minor subject headings (MeSH, National Library of Medicine Medical Subject Heading), TI title, AB abstract, and SU subject.

To ensure relevant publications were not overlooked, studies that met the inclusion criteria were hand-searched. A forward citation search was performed using Google Scholar to find studies not yet indexed in the aforementioned databases [28].

Selection process

The following criteria were used to determine inclusion:

1. Peer-reviewed studies.

2. Studies published in the English language.

3. Studies published between 1 January 2016 and 12 April 2024.

4. For the first aim – studies must include cohorts with dysvascular PFA, irrespective of the level of PFA.

5. For the second aim – studies must include separate cohorts with dysvascular PFA and TTA

6. Studies that measure proportionate mortality (or present data on survival from which proportionate mortality can be calculated) regardless of the time point at which it was measured following amputation.

Studies that group together people with different levels of lower limb amputation (e.g., PFA and TTA) or causes (e.g., trauma and PVD) were excluded, along with editorials, letters, conference abstracts, and opinion pieces. The International Standards Organisation (ISO) definitions of PFA and TTA were used and therefore, all levels of PFA (including toe amputation) were included [32]. In keeping with the ISO definition, ankle disarticulation (i.e., Syme amputation) was not considered a PFA and as such, studies regarding this amputation level were excluded [32]. Given the ambiguity of terms such as “minor” or “major” amputation, we have not used those terms, except when reporting the results of other studies where those collective nouns were used. In those instances, we have described the level of amputation included using ISO terms.

Search results were exported directly to an EndNote 20 (Clarivate, Philadelphia, PA) library and duplicates removed. The primary investigator (Z.W.) screened the search results based on a review of the title, abstract, or full-text study as necessary. An additional opinion was sought from a second investigator (M.D.) in cases of doubt, with disagreements resolved through discussion until consensus.

Data collection process

Based on the Cochrane Consumers and Communication Review Group’s data extraction template, [33] data from included studies were systematically recorded in an Excel (Microsoft Corporation Inc., Redmond, WA) spreadsheet including: socio-demographic, methodological, proportionate mortality, and quality appraisal details (Additional File 3). This spreadsheet was developed, tested, and used in a previous systematic review [5].

For each of the included studies, data were initially extracted by one investigator (either E.R. or Z.W.) before being independently verified by a second investigator (either M.D. or E.R.). Inconsistencies in data entry were resolved through discussion until consensus. If additional information or clarification was required, the authors of the original research were contacted via email with one follow-up a week later, as needed.

Study risk of bias assessment

The McMaster Critical Review Forms were used to assess methodological quality and identify sources of bias [34]. The appraisal tool was appropriate for use with a wide variety of study designs and included guidelines to reduce the likelihood of errors [35–37]. Results from the McMaster Critical Review Form were recorded in the Excel spreadsheet with detailed comments to support checklist items (Additional File 3).

Statistical analyses

Random effects proportionate meta-analyses (StatsDirect, Wirral, UK) were conducted to synthesise proportionate mortality at discrete time points following amputation. All studies were included in the proportionate meta-analyses as preserving information regarding the heterogeneity of results was important to explore underlying causes of variation and uncertainty in the point estimates reported [38,39]. Results were reported using a forest plot along with the point estimates and 95% confidence intervals. Relative risk meta-analyses were also undertaken to compare the RR of mortality between the PFA and TTA cohorts. Heterogeneity of the results between studies was quantified using the I² statistic and explored using the findings of the risk of bias assessment. Meta-analyses were accompanied by a narrative review to explain the variations between studies using the findings from the risk of bias and extracted data.

Results

Study selection

The database search yielded 7487 studies (Figure 1). Once duplicates were removed, 3320 studies were screened against the inclusion criteria based on the title and abstract. Forty-three full-text studies were examined, of which 18 met the inclusion criteria. The forward citation search yielded two additional studies that both met the inclusion criteria. Of the 20 studies that met the inclusion criteria, two were excluded due to significant methodological or reporting issues including: inconsistent description of number of people included [9] or the time period over which data were collected [40]. One study was excluded given “minor” amputations were not operationally defined and we were unable to confirm the included amputation levels after efforts to contact the authors [41]. Therefore, 17 studies were included in this review (Figure 1).

Figure 1. PRISMA flowchart.

Study characteristics

The majority of included studies reported outcomes from countries with well-developed healthcare systems including the USA [11,12,42,43], Malta [44], China [45,46], Italy [13], Denmark [47], UK [15,48], and Spain [49].

Given the inclusion criteria, all the included studies reported proportionate mortality following dysvascular PFA [11–15,42–53]. One study also presented proportionate mortality for separate PFA and TTA cohorts [53]. The time points at which proportionate mortality was reported have been summarised in Table 2.

Table 2. Summary of study designs and outcomes of included studies.

Author	Study design		Outcomes
			Proportionate mortality partial foot amputation					Proportionate mortality transtibial amputation
	Retrospective	Prospective	30 day	1-year	3-years	5-years	Other^a	3day	1-year	3-years	5-years	Other^a
Adams et al. [12]	X		X		X
Gurney et al. [14]	X			X	X		X
Vassallo et al. [44]		X		X			X
Zhang et al. [46]		X		X	X		X
Norvell et al. [53]	X			X					X
Lepow et al. [42]	X			X			X
Chu et al. [45]		X		X	X	X
Faglia et al. [13]	X			X	X		X
Wilbek et al. [47]	X					X
Joyce et al. [15]	X			X	X	X	X
Chung et al. [52]	X			X
López-Valverde et al. [49]	X			X	X	X	X
Blanchette et al. [50]	X			X
Varghese et al. [43]	X			X	X	X	X
Cascini et al. [51]	X			X			X
Birmpili et al. [48]	X			X	X	X
Rolle et al. [11]	X			X

^a Studies that reported proportionate mortality at other time points (e.g., 6-year post-PFA).

The majority of the included studies were retrospective cohort designs (Table 2) from single-centres [12,13,15,42–47,49,52] and as such, they generally contained smaller samples when compared to studies of hospital network [11,51] or national-level [14,48,53] data sets (Table 3).

Table 3. Summary demographic characteristics of included studies.

Author	Partial foot amputation				Transtibial amputation				Summary of demographic characteristics
	Total no. of PFA (N)	Percent w/diabetes (%)	Sex (% male)	Age (mean ± SD years)	Total no. of TTA (N)	Percent w/diabetes (%)	Sex (% male)	Age (mean ± SD years)	A	B	C	D	E	F	G	H	I	J	K	L	M	N	O	P	Q	R
Adams et al. [12]	375	87%	70.90%	66.4 (12.8)					X		X	X	X						X	X	X		X	X
Gurney et al. [14]	1217	100%	68%	65 (13.3)					X	X									X				X	X
Vassallo et al. [44]	81	100%	72%	70 (12.3)							X		X	X			X		X	X
Zhang et al. [46]	97	100%	63.99%	66.5 (9.9)									X	X			X		X	X	X	X	X
Norvell et al. [53]	1504	Not reported for PFA or TTA cohorts. For the total group (PFA, TTA, and TFA), 62% had diabetes.	Not reported for PFA or TTA cohorts. For the total group (PFA, TTA, and TFA), 99.1% were male.	Not reported for PFA or TTA cohorts. For the total group (PFA, TTA, and TFA), mean age was 68 ± 11 years.	3261	Not reported for PFA or TTA cohorts. For the total group (PFA, TTA, and TFA), 62% had diabetes.	Not reported for PFA or TTA cohorts. For the total group (PFA, TTA, and TFA), 99.1% were male.	Not reported for PFA or TTA cohorts. For the total group (PFA, TTA, and TFA), mean age was 68 ± 11 years.	X	X	X	X	X			X		X	X	X		X	X
Lepow et al. [42]	41	75.60%	65.80%	65.6 (11.1)							X		X		X	X			X	X	X		X	X	X
Chu et al. [45]	245	100%	59%	69.27 (9.39)							X	X		X			X		X				X	X		X
Faglia et al. [13]	83	100%	69%	71.4 ± 9.3									X				X		X			X			X	X
Wilbek et al. [47]	777	100%	77%	65.2 (no SD)							X	X	X	X			X		X			X		X		X
Joyce et al. [15]	47	100%	85%	60 (no SD)							X			X					X					X	X
Chung et al. [52]	49	Not reported for the PFA cohort. For the total group, 72% had diabetes.	Not reported for the PFA cohort. For the total group, 57.6% were male.	Not reported for the PFA cohort. For the total group, the mean age was 71.2 (no SD).													X			X				X	X
López-Valverde et al. [49]	116	Not reported for the PFA cohort. For the total group, 83.3% had diabetes.	Not reported for the PFA cohort. For the total group, 78.3% were male.	Not reported for the PFA cohort. For the total group, the median age was 74 (Q1: 64, Q3: 81).							X				X		X	X		X			X	X	X	X
Blanchette et al. [50]	70	100%	85.70%	57.4 (11)					X		X		X			X		X		X	X	X
Varghese et al. [43]	389	78.10%	67.10%	63.3 (14.2)					X		X	X			X		X			X		X		X		X
Cascini et al. [51]	736	100%	73.20%	Mean and SD not reported. Age reported as % age groups 35–54, 55–64, 65–74, and 75+						X								X	X	X				X	X
Birmpili et al. [48]	22 188	83.52%	75.9%	Mean and SD not reported. Age reported in ranges of 40–59, 60–69, 70–79, and >80															X				X			X
Rolle et al. [11]	373	77.6%	67.5%	56.8 (12.1)					X		X				X	X					X		X	X

A: race/ethnicity; B: socio-demographic details (education level, relationship status, employment status, residential status, geographic region, and economic status); C: smoking/tobacco use; D: other lifestyle behaviours (alcohol use, malnutrition, and obesity); E: exams/tests (physical exams, lab tests, and cultures); F: Hb1Ac; G: other biomarkers (GHb, hyperlipidaemia); H: BMI; I: sensation (neuropathy, impaired protective sensation, and vibration perception threshold); J: mental health (major depressive disorders and post-traumatic stress disorder); K: comorbidities (Charlson Comorbidity Index, frequency of comorbidities, Anesthesiology Association of America Physical Status Scale); L: blood pressure/flow issues (missing pulses, Ankle Brachial Index); M: coronary artery disease; N: stroke/CVA; O: other cardiovascular issues (cardiovascular disease, myocardial infarct, heart attack, peripheral arterial disease, peripheral vascular disease, and chronic obstructive pulmonary disease); P: renal issues (nephropathy, renal dialysis, chronic renal failure, renal insufficiency, and end stage renal disease); Q: prior amputation; R: other (immunosuppressant use, joint replacement, Charcot collapse, footwear type, perceived premorbid health, retinopathy, ulcer, and traumatic brain injury).

Most studies reported proportionate mortality for a discrete level of PFA such as: transmetatarsal [12,15,42,46,53], toe [44,45,50], Chopart [13], or partial first ray [50]. Two studies grouped toe and ray amputations together [11,52]. Five studies did not stratify their data by level of PFA and instead described the cohort using the collective noun “minor amputation” [14,43,47–49,51] which included different levels of PFA. While some studies also included people with TTA [14,43,49,51–53], most grouped people with TTA into a single “major amputation” cohort along with people living with transfemoral amputation; the exception reported proportionate mortality for the TTA-only cohort [53] in keeping with the inclusion criteria. None of the included studies were designed to compare proportionate mortality between cohorts based on level of amputation.

Given the population of interest, the participants in the included studies were typically older males living with diabetes and other comorbidities (Table 3).

Quality appraisal/risk of bias

There were several recurrent issues that impacted both the internal and external validity of the included studies. Each of these issues has been discussed in the paragraphs that follow.

While most studies reported age, sex, and the presence of diabetes, more detailed information about the participants was often poorly reported which made it difficult to generalise these findings to the wider population of people living with dysvascular PFA or TTA. For example, sociodemographic factors such as race/ethnicity, or the presence of common comorbidities known to affect mortality (e.g., the presence of coronary arterial disease) were inconsistently reported (Table 3). Similarly, the type of diabetes was often not reported despite the fact that proportionate mortality differs for cohorts with type 1 and type 2 diabetes [47,54].

Cointervention likely affected all included studies given the healthcare required to manage comorbid conditions such as end stage renal disease or coronary heart disease (Table 3). Looking specifically at coronary heart disease as an illustrative example, common interventions such as smoking cessation, or the use of beta blockers to manage hypertension, are known to affect mortality outcomes.

There were several issues with the design of studies that reduced confidence in the proportionate mortality reported. Ten studies retrospectively measured mortality in a discrete time window (e.g., 5-year post-amputation) and used that data to estimate proportionate mortality from Kaplan–Meier’s analyses [13,14,42,43,45–49,51]. Given this approach, these studies likely included many data points that underpinned the 1-year proportionate mortality estimate, and comparatively few data points for the 5-year mortality estimate, increasing uncertainty. None of these studies reported an a priori sample size calculation to justify the number of participants required for the Kaplan–Meier analysis at these different time points, nor did they report confidence intervals alongside the Kaplan–Meier estimates to showcase uncertainty of the estimated proportionate mortality in the years after amputation as participants were censored (i.e., died, or lost to follow up and removed from the analysis).

Proportionate mortality

Given the inclusion criteria, all included studies reported either all-cause proportionate mortality (or survival from which mortality could be calculated) and included additional information regarding a discrete time point at which mortality was measured (e.g., 1-year after amputation).

Proportionate mortality was reported either by count of the sample that had died at each time point [11,12,15,44,50,52,53] or by using a Kaplan–Meier analysis to generate an estimate [13,14,42,43,45–49,51]. All studies reported proportionate mortality at 1-, 3-, or 5-years post-amputation [11–15,42–53]. There were also isolated data for other time points post-amputation including: 30 days [12]; 1-, 3-, 6-, and 9-months [15,44,46]; or 2-, 4-, 6-, 7-, and 8-years [13,14,42,43,51].

Consistent with the method we adopted from a previous systematic review [5], we reported proportionate mortality for 30 days, and 1-, 3-, and 5-years post-amputation.

Proportionate mortality following dysvascular PFA

For 30 day mortality, one study [12] reported that 2.1% (no confidence interval) of individuals died within 30 days after PFA (specifically, transmetatarsal amputation). Given this baseline, proportionate mortality increased from 1-year (13.9%, 95%CI 11.6–16.3), 3-years (30.1%, 95%CI 25.1–35.4), and 5-years (42.2%, 95%CI 36.0–48.5) following PFA (Figure 2). There was considerable heterogeneity in the results between studies (I² >80%, p < 0.01). While the heterogeneity between studies did not detract from the precision of the point estimate 1-year post-PFA, uncertainty increased over time (Figure 2). Studies that contributed most to the heterogeneity received less weighting in the random-effects meta-analysis and thereby had less influence on the point estimate.

Figure 2. Proportionate mortality at 1-, 3-, and 5-years after partial foot amputation (random effects meta-analysis).

As measures of proportionate mortality do not control for variations in the sample, such as older age or the presence of comorbidities, these were a source of variation in the results between studies. For example, in the study by Faglia et al. [13] proportionate mortality at 1-year post-amputation was relatively high (36.0%, 95%CI 26.0–47.0%) compared to other studies (Figure 2); likely because the sample was older (71.4 ± 9.3 years), had diabetes for many years (19.3 ± 10.9 years), included a large proportion with failed previous amputations (42.5%), had significant wounds characterised by exposure of the bone or joint with infection and ischaemia (69.9%, Texas wound classification D3), and required emergency surgery due to gangrene or osteomyelitis (65.9%). As such, it is not surprising that proportionate mortality was greater when compared to other studies with more representative samples. By contrast, Chu et al. [45] reported a much lower proportionate mortality at 1-year post-amputation (6.0%, 95%CI 3.0–9.0%, Figure 2). While participants in this investigation were similar in age (69.27 ± 9.39 years) to Faglia et al. [13], it is likely they had less advanced systemic disease given that the time with diabetes was shorter (weighted mean 12.8 ± 7.3 years) and that toe- or metatarsophalangeal-amputation was clinically indicated [45].

No studies reported proportionate mortality stratified by different levels of PFA.

Proportionate mortality following dysvascular PFA compared to TTA

One study reported proportionate mortality 1-year after PFA (17.7%) and TTA (24.7%) [53]. In comparison to the PFA cohort (specifically, transmetatarsal amputation), the TTA cohort had a significantly greater RR of dying at 1-year post-amputation (RR 1.5, 95%CI 1.4–1.7, p < 0.0001). Personal correspondence with the authors confirmed that the participants in this study [53] were the same as another cohort study [55] with the exception of a very small proportion who were excluded in the later publication [55] due to constraints of the regression modelling. Based on the demographic and health information for the PFA and TTA cohorts [55], it was possible to discern that, while the PFA and TTA cohorts were similar in terms of age (PFA 65.6 ± 10.0, TTA 66.8 ± 10.2 years), proportion with diabetes (PFA 82.9%, TTA 77.8%) and sex (99.1% male for both PFA and TTA), the PFA cohort had a higher proportion of people with normal Ankle Brachial Index compared to the TTA cohort (PFA 47.0%, TTA 39.8%). The PFA sample also had a smaller proportion with chronic obstructive pulmonary disease (PFA 13.3%, TTA 18.1%) and any revascularisation in the past year (PFA 35.8%, TTA 41.9%). Finally, the PFA sample also showed a higher proportion with independent physical function (PFA 65.0%, TTA 56.5%). These data suggest that the individuals in the TTA cohort had worse systemic health at the point of amputation compared to the PFA cohort, which likely contributed to the higher mortality in the years that followed.

Discussion

Given the aim of this systematic review, the following discussion has been subdivided to interpret the proportionate mortality results for 30 days, 1-, 3-, and 5-years following dysvascular PFA, and compare these outcomes between cohorts with PFA and TTA. The final subsections of the discussion focus on the clinical relevance and limitations of this review.

Proportionate mortality following dysvascular PFA

Following PFA, proportionate mortality increased over time from 2.1% (30 days), 13.9% (1-year), 30.1% (3-years), and 42.2% (5-years). These results are similar to other systematic reviews [5,56] that reported proportionate mortality at these same time points (Table 4). The similarities in findings between these reviews reflect, to some extent, the similarity of method designs. For example, these systematic reviews [5,56] focused on original research of people with dysvascular amputation and used a random-effects meta-analysis to synthesise data at the same timepoints. These reviews [5,56] also included many of the same original studies, further explaining the similarities given the time periods of included studies overlapped.

Table 4. Proportionate mortality reported in different systematic reviews.

	Proportionate mortality following PFA
Systematic reviews	30 days %, (95%CI)	1-year %, (95%CI)	3-years %, (95%CI)	5-years %, (95%CI)
Yammine et al. [56]	3.5 (2.7–4.4)	20.0 (15.2–25.0)	28.0 (24.5–31.6)	44.1 (47.1–47.6)
Dillon et al. [5]	2.8 (2.2–3.6)	17.3 (14.6–20.2)	30.6 (23.0–38.7)	41.2 (32.6–50.1)
This review	2.1 (No CI)	13.9 (11.6–16.3)	30.1 (25.1–35.4)	42.2 (36.0–48.5)

CI: confidence interval.

The 95%CI reported by Yammine et al. [56] have been multiplied by 100 to match the point estimates reported.

While the findings in this study were very similar to previous systematic reviews [5,56], there were some noteworthy differences. First, in this study, proportionate mortality 30 days after PFA was lower than the 95%CI reported in other systematic reviews (Table 4). Given that just one included study reported a 30-day mortality rate [12], we suggest that the results from the previous systematic reviews [5,56] provide a more reliable estimate of 30-day proportionate mortality as they included several studies that allow both a point estimate and 95% confidence intervals to be calculated.

Second, while the point estimates and confidence intervals were similar across these systematic reviews (Table 4), we do not have confidence in the 95% confidence interval reported by Yammine et al. [56] at 5-year post-PFA given the point estimate sits outside the confidence interval (44.1, 95%CI 47.1–47.6), and the confidence interval seems narrow given the heterogeneity reported (I² = 88.7%). Finally, while previous systematic reviews [5,56] have used similar databases, search terms, and inclusion criteria, there were differences in the choice of critical appraisal checklist and, more importantly, how these data were used to explain the underlying causes of variation between studies. One of the strengths of this review was that the results narrative synthesised observations from the critical appraisal (i.e., McMaster appraisal checklist) and extracted data (e.g., patient demographics) to explain variation between studies, such as the impact that older age or comorbid health conditions had on mortality outcomes. The ability to explain the underlying causes of variation between studies is important to engender confidence in the point estimates reported and support clinicians to apply these findings to a diverse population.

Proportionate mortality following dysvascular PFA compared to TTA

Based on the results of the one included study [53] that reported proportionate mortality 1-year after dysvascular PFA (17.7%) and TTA (24.7%), the RR of dying in the first year after TTA was 50% higher than after PFA (RR 1.5, 95%CI 1.4–1.7). This observation was similar to a previous systematic review (RR 1.5, 95%CI 1.4–1.6) that included numerous studies given the time window for inclusion included all indexed studies up to 31 December 2015, and therefore we can have confidence in the finding despite only having one included study [12].

Studies that have used appropriate statistical techniques that control for interactions between factors that confound proportionate mortality (e.g., older age, higher amputation level, presence of comorbid disease such as end-stage renal disease or coronary arterial disease) have reported that amputation level does not independently affect mortality [6–8]. Therefore, the increased RR of dying in the years following TTA (as compared to PFA) may be due to the more advanced systemic disease that made the TTA a preferred amputation surgery, and has little to do with the choice of amputation level per se [6–8].

Given this systematic review used the same method as a previous systematic review [5], together they provide continuous data from the beginning of indexed databases to 12 April 2024. Meta-analyses based on aggregated results show the proportionate mortality after PFA increased over time from 30 days (2.8%, 95%CI 2.2–3.4), to 1-year (15.3%, 95%CI 13.8–17.0), 3-years (30.3%, 95%CI 26.5–34.2), and 5-years (41.5%, 95%CI 35.9–47.2) as depicted in Figure 3. Similarly, in comparison to the PFA cohort, the TTA cohort had a significantly greater RR of dying at 1-year post-amputation (RR 1.45, 95%CI 1.36–1.54, p < 0.0001).

Figure 3. Updated proportionate mortality at 1-, 3-, and 5-years after partial foot amputation (random effects meta-analysis) including data from this review and preceding review [5].

Clinical relevance

Proportionate mortality following dysvascular PFA has not changed significantly in recent years given the results of this systematic review are comparable to a previous systematic review that included literature published on 31 December 2015 [5]. Given this finding, the proportionate mortality reported in patient decision aids [5,25,57] is accurate and as such, only a minor revision is necessary to update any supporting references based on this review.

Future research

In recent years, the volume of literature reporting proportionate mortality following dysvascular PFA and PFA/TTA has grown substantially compared to a previous systematic review [5]. However, many of the new studies are from small centres with a small number of participants and do not provide a compelling rationale that would justify further research reporting proportionate mortality. We suggest that further research of isolated healthcare services with small patient numbers will not add value to the already well-developed body of literature on this topic unless there are significant changes in clinical practice that have potential to improve mortality outcomes that warrant publication [58].

While this systematic review focused on proportionate mortality, there are many other outcomes that are also important to help inform difficult decisions about amputation surgery including: mobility, health-related quality of life, delayed wound healing, or the risk of reamputation, among others [5,25,57]. There are opportunities to review literature on these topics to help ensure current evidence has been appraised and synthesised to provide the data needed to inform surgical consultations or patient-facing decision aids.

Limitations

The results of this investigation should be considered in light of a number of limitations.

We acknowledge that proportionate mortality has been reported for other timepoints not included in this review such as 3-, 6-, and 9-months [15,44,46]; or 2-, 4-, 6-, 7-, and 8-years [13,14,42,43,51]. Since we adopted the method of a previous systematic review [12], we used the same timepoints to report the outcomes given they are the most commonly reported, and adequately demonstrate changes in mortality and RR over time. For those wanting to utilise data from other time points, or look at longer term mortality outcomes, these data have been reported in Additional File 3.

Given the multiple timepoints at which proportionate mortality has been reported, the result narratives deliberately use illustrative examples to showcase the risks of bias and the impact on the results, rather than reporting risk of bias for each study. Full details of the risk of bias assessment are included as Additional File 3.

Conclusions

This systematic review highlights that proportionate mortality following dysvascular PFA, or PFA and TTA, has not changed in recent years, despite an increase in new research. As such, outcomes of proportionate mortality reported in a previous systematic review and patient-facing decision aids reflect contemporary evidence.

Acknowledgements

We would like to acknowledge Professor Kate Webster of La Trobe University for producing the meta-analyses in this review.

Disclosure statement

Professor Michael Dillon, Dr Matthew Quigley, and Professor Stefania Fatone were the developers of a patient-facing decision aid (i.e., Amputation Decision Aid) [57]. No financial support was received for the current systematic review. The authors declare no other competing interests.

Data availability statement

The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials.

Additional information

Funding

None.