1 Introduction

Numerous factors influence the specificity of muscle biopsy. Pathology can affect muscle in a patchy distribution, and consequently, sampling error can occur [1]. Although techniques to assess muscle pathology have evolved to include a growing number of specialized stains and genetic tests that supplement traditional methodologies [2], findings may still be non-specific.

In the past, muscle biopsy was a fundamental investigation to diagnose the idiopathic inflammatory myopathies (IIM) [3, 4]. The IIM subgroups polymyositis (PM), dermatomyositis (DM), immune-mediated necrotizing myopathy (IMNM), inclusion body myositis (IBM), anti-synthetase syndrome (ASyS) and overlap myositis (OM) have distinguishing histological features [5, 6]. With the discovery of myositis-specific antibodies (MSA), diagnosis is now possible in some situations without the need for muscle biopsy. However, biopsy confirmation of IIM remains important, especially in the absence of MSA [7], with suspicion of a non-inflammatory myopathy [8], and for the diagnosis of IBM [9].

Rarely, a patient may undergo a repeat biopsy. Sometimes, repeating a biopsy is prompted by an inadequate or inappropriate specimen being obtained from the initial biopsy. Other reasons might include an unclear diagnosis, a change in patient presentation suggesting new pathology, or a lack of response to treatment prompting reconsideration of a prior diagnosis [10]. Generally, muscle biopsies are not used for monitoring treatment response.

The indications, clinical utility, and shortcomings of a single muscle biopsy are well-recognized. However, there is sparse data evaluating the utility of repeating a muscle biopsy [10, 11]. Additionally, the predictors of a repeat biopsy contributing to clinical decision-making and the role of repeating a biopsy in the setting of non-specific histological abnormalities have not yet been explored. We aimed to describe the indications and outcomes of repeating a muscle biopsy.

2 Methods

In South Australia (SA) between 2000 and 2023, muscle biopsies were performed by one of two methods: a core biopsy using a Bergström needle performed by a specially trained rheumatologist or neurologist, or an open surgical biopsy by a surgeon.

In SA, all muscle biopsies are evaluated at a single laboratory, SA Pathology. Samples are subdivided for formalin and glutaraldehyde fixation, and for snap freezing. A variety of stains are applied including hematoxylin/eosin, modified Gomori trichrome at pH 3.4, ATPase at pH 4.6, and COX/SDH histochemistry. Immunostaining for CD45 (identifying lymphocytes) and CD68 (identifying macrophages), MHC-I and MHC-II, neonatal myosin, dysferlin, caveolin 3, and p62 are routinely performed. All biopsies are further examined ultrastructurally using electron microscopy.

The SA Pathology laboratory uses established histological criteria to diagnose IIM and its subsets [5, 6]. The entity of myositis not-otherwise-specified (MNOS) is additionally recognized; this diagnosis is applied when there are features of skeletal muscle inflammation albeit falling short of satisfying histological criteria for a particular subset of IIM. The term “non-specific myopathy” is applied when there are some, often milder, histological aberrances within muscle which appear non-inflammatory and do not satisfy criteria for an alternative diagnosis, such as a muscular dystrophy.

The slides and reports of all SA muscle biopsies have been systematically stored at SA Pathology for more than 30 years. All biopsies have been reported by one of six experienced neuromuscular pathologists and are subject to regular peer review.

We reviewed muscle biopsy reports over a 24-year period (2000–2023) and identified pediatric and adult patients who had undergone more than one muscle biopsy. Patients who had a repeat muscle biopsy because of an inadequate specimen (e.g., insufficient muscle tissue for analysis) were excluded.

For each patient, the following information was extracted from the clinical records: demographics, indication for repeat muscle biopsy, and immunomodulator exposure between biopsy one (B1) and biopsy two (B2). From the pathology records, the following data were documented: site and method of muscle biopsies, pathological diagnosis as assigned by the reporting pathologist, presence/absence of histopathological features (necrosis, regeneration, endomysial inflammatory infiltrate, interstitial inflammation, perifascicular atrophy, fiber atrophy, fiber hypertrophy, vacuoles [non-rimmed and rimmed]), immunohistochemical features (proportion of COX-/SDH+ fibers, CD45 and CD68 staining, MHC-I and MHC-II upregulation) and ultrastructural abnormalities (tubulofilamentous inclusions and mitochondrial defects).

The indications for repeat biopsy were determined and considered amongst three broad groups: (1) those in whom clinical features had raised uncertainty about the initial histological diagnosis (B1), thus prompting a search for an alternative or confirmatory pathological diagnosis, (2) those patients whose clinical trajectory was unexpected, and (3) patients in whom further tissue for specialized testing or a more readily interpretable muscle source was requested by the pathologist.

To assess the change in histological diagnosis from B1 to B2, diagnoses were sub-classified into non-specific diagnoses (non-specific myopathy and MNOS) and specific diagnoses (PM, DM, IMNM, IBM, genetically determined myopathy, denervation, atrophy [generalized and not restricted to a specific fiber subtype], vasculitis and normal). Broadly, two outcomes were assessed in comparing serial biopsies: Outcome 1 was defined as a non-specific diagnosis (B1) becoming a specific diagnosis (B2), and Outcome 2 was defined as a change in specific diagnosis between B1 and B2. Histological and immunohistochemical changes were also documented.

For patients meeting Outcome 1 or 2, the clinical significance of meeting these histopathological outcomes was assessed through reviewing the clinical records, where available, for a change in, or confirmation of, clinical decision making.

Descriptive statistics were applied to this retrospective cohort. Odds ratios (OR) were calculated for predictors of meeting Outcome 1 or 2 using logistic regression. Fisher's exact test was used for the comparison of categorical variables. Sankey plots were generated to demonstrate the change in biopsy approach and diagnosis between biopsies. Stata version 18 (StataCorp, College Station, TX) and the sankey v1.74 package were used.

This research was conducted according to the 1964 Declaration of Helsinki and all later amendments. The retrospective collection of clinical and pathological data for this study was approved by the Central Adelaide Local Health Network Human Research Ethics Committee (reference number 20599).

3 Results

From 2000 to 2023 inclusive, a total of 3089 biopsies were assessed at SA Pathology. Of these, 112 (3.6%) were repeat biopsies performed on 107 patients, five of whom had three muscle biopsies. Table 1 demonstrates the demographic, biopsy, and histological characteristics of this cohort. Although most patients were adults, three were pediatric at the time of B1, two of whom underwent repeat biopsy before the age of 18 years. The number of repeat muscle biopsies peaked in 2010–2014 (33 repeat biopsies) (Figure S1).

The proportion of biopsies performed using an open technique progressively increased with each serial biopsy. Few biopsies were specifically targeted to abnormal signal on muscle MRI. B2 more commonly occurred in the same muscle on the contralateral side (56/107) than in a different muscle (32/107) or the same muscle on the same side (19/107). Although the vastus and gastrocnemius muscles were most frequently targeted, open biopsy of other muscles (e.g., deltoid, extraocular muscles and paraspinal muscles) also occurred.

Amongst patients undergoing repeat muscle biopsy, the most frequent initial histological diagnosis was non-specific myopathy followed by MNOS (Table 1). The Sankey plot in Figure S2 shows the change in histological diagnoses between B1 and B2 in more detail. More than half with an initial diagnosis of a non-specific myopathy (21/34, 61.8%) changed to a specific diagnosis in B2 (Outcome 1). Most patients with an initial diagnosis of one of the IIM subtypes (excluding IBM) remained within the IIM group (21/39, 53.8%) with four (10.3%) being reclassified as IBM. Half of the 16 biopsies initially diagnosed as MNOS were reclassified on subsequent biopsy: eight satisfied criteria for a more specific IIM subtype (PM [2], DM [1], IMNM [3], IBM [2]), three were diagnosed with non-specific myopathy, and one developed atrophy as a primary diagnosis without inflammation.

Five patients had a diagnosis of IBM on B1, all of whom retained this diagnosis on B2. In three of these patients, repeat biopsies were undertaken for diagnostic confirmation in the setting of an atypical pattern of weakness. Although IBM is considered refractory to immunosuppressive therapies, it appeared that the reason for repeat biopsy in the remaining two patients was a lack of response to immunosuppression. Of the seven patients who did not have IBM on B1 but subsequently had this diagnosis on B2, all had new rimmed vacuoles on B2, and three had new mitochondrial changes suggestive of IBM.

Of the eight patients with PM on B1, one was diagnosed again with PM on B2, two with IBM, one with MNOS, one with muscular dystrophy, and three with non-specific myopathy.

Seven patients had a primary histological diagnosis of muscle atrophy on B2. Among these seven patients, only one had atrophy on the previous biopsy; the atrophy had evolved between biopsies in three patients with IIM (two with IMNM and one with MNOS, all on immunosuppressive therapies) and three patients with non-specific myopathy (one of whom was on immunosuppressive therapy) on B1.

Overall, a substantial proportion of repeat biopsies (48/112, 42.9%) met Outcome 1 (31/48) or Outcome 2 (17/48) (Table 1). The altered histological diagnosis on B2 or B3 clearly guided clinical decision making in 17/22 (77.3%) patients in whom detailed medical records were available for review. These decisions included: addition of further immunomodulator (6/12), decrease/cessation of immunosuppression (6/12), referral for genetic testing (2/12) and initiation of an exercise/rehabilitation plan (3/12).

Sampling by an open surgical approach for B2 led to a higher proportion of biopsies meeting Outcome 1 or 2 than when repeat sampling was performed by needle biopsy (open: 41/81, 50.6% vs. needle: 7/31, 22.6% biopsies meeting Outcome 1 or 2, p < 0.01) (Figure S3).

The indications for a repeat muscle biopsy were for a specific diagnostic query (36/112, 32.1%) or unexpected clinical trajectory (74/112, 66.1%) in the majority; only two patients underwent repeat biopsy for a “pathological” indication (2/112, 1.8%) (Table 2). The most common diagnostic queries were consideration/confirmation of IBM or a genetically determined myopathy. With a high suspicion of IBM before B2, more than half (10/13) repeat biopsies confirmed this entity.

Unexpected clinical trajectory as an indication for repeat biopsy was predominantly driven by symptoms (51/74, 68.9%) or a lack of expected response to treatment (17/74, 23%). In situations of diagnostic query, a repeat muscle biopsy met Outcome 1 or 2 in 16/36 (44.4%) of cases, and with unexpected clinical trajectory, 32/74 (43.2%) of cases.

We next sought to determine which features may predict the greatest likelihood of achieving Outcome 1 or 2 with repeat biopsy (Table S1). On univariate logistic regression, the only variable with a significant odds ratio was the use of an open biopsy method for repeat biopsy (OR 4.3 [IQR 1.6–11.5], p = 0.004). When adjusted for the other variables, however, biopsy method was collinear.

Forty-five (42.1%) repeat biopsies were performed following exposure to immunomodulators. Prednisolone was the most common exposure (38/45, 84.4%) followed by methotrexate (21/45, 46.7%), azathioprine (7/45, 15.6%), hydroxychloroquine (7/45, 15.6%), intravenous immunoglobulin (6/45, 13.3%), mycophenolate mofetil (5/45, 11.1%) and sulfasalazine (3/45, 6.7%).

Patients with a non-specific myopathy on B1 were less likely to be treated with immunomodulators than the rest of the cohort (8/35 vs. 37/77, p = 0.01). However, patients with MNOS on B1 were treated with immunomodulators as often as those with a more specific non-IBM IIM diagnosis (11/18 vs. 16/23, p = 0.7).

We sought to determine the evolution/attenuation of histopathological features of inflammation over time. Table 3 documents the changes between B1 and B2 of features considered to be indicative of inflammation and autoimmunity, namely the presence of an inflammatory infiltrate (endomysial and/or perimysial), immunohistochemical staining for CD45 and/or CD68, and MHC-I/II upregulation. A minority of cases demonstrated improvement in inflammation. Almost a quarter of repeat biopsies (24/112, 21.4%) showed new evidence of inflammation, some despite immunomodulatory therapies (10/24, 41.7%). IBM and PM accounted for about half of patients showing greater MHC-I and/or MHC-II expression on repeat biopsy (new MHC-I: 6/14; MHC-II: 5/13 with IBM). Of those with new cellular infiltrates as indicated by CD45 and CD68 staining, 9/14 had concurrent MHC upregulation, but only 3/14 had elevated CK.

4 Discussion

Although repeat biopsies represented a small portion of muscle biopsies in South Australia, clinical decision making was impacted in the majority of cases in which follow-up data was analyzed. Based on our findings, repeating a biopsy in patients with a non-specific myopathy and those in whom IBM or a genetic myopathy is suspected has the highest yield. The larger samples retrieved with open sampling led to the highest impact.

Moreover, beyond the diagnostic and treatment implications of our findings, this study provides some important insights into the evolution of muscle pathology: (1) myofibre atrophy may develop in non-IBM IIM patients despite immunosuppressive therapies and also, importantly, in patients with non-specific myopathies, suggesting that non-specific changes may still indicate muscle pathology with the potential to evolve into damage, therefore warranting further early investigation and/or treatment, and (2) MHC-I and II expression can be attenuated with treatment.

Repeating a muscle biopsy in our cohort was predominantly motivated by a specific diagnostic consideration or due to unexpected clinical progress. For the former indication, queries of IBM and genetic myopathies more frequently resulted in a specific diagnosis being discovered on repeat biopsy. Although IBM has clinical features distinct from other subsets of IIM, demonstrating myopathology on muscle biopsy remains a required component of diagnostic criteria for IBM [9]. As some of the salient histological features of IBM may take time to evolve and can be missed on initial biopsy, it is not surprising that repeating muscle biopsy with a pre-test clinical suspicion of IBM more frequently leads to a specific histological diagnosis.

Notably, however, all patients with IBM on initial muscle biopsy had the same diagnosis on repeat biopsy, suggesting that there is no utility in repeating a biopsy once IBM has already been diagnosed. With our current understanding of the refractory nature of IBM, the two repeat biopsies performed in patients with IBM on B1 due to a lack of response to immunosuppression had an expectedly low yield.

That genetic myopathies were as common a diagnostic query as IBM in our cohort may be a consequence of analyzing a historical cohort extending back 24 years. In the present day, genetic myopathies are more frequently diagnosed with increasingly sensitive tests of peripheral blood DNA, potentially avoiding the need for a muscle biopsy [12].

Another consideration may also be recent shifts in understanding of IIM classification. That only one of eight patients with PM diagnosed on B1 retained this diagnosis on B2 is consistent with PM now being considered a much rarer subtype than in the past [13]. Our finding that many with an initial diagnosis of PM subsequently were re-classified as MNOS or non-specific myopathy likely reflects attenuation of inflammation with treatment.

In our cohort, more than half of patients with an initial biopsy reported as a non-specific myopathy had a more specific finding evident on repeat biopsy, supporting the dynamic nature of some muscle pathologies over time. That three patients with a non-specific myopathy on initial biopsy subsequently developed histological atrophy (clinical atrophy notably being a sign of muscle damage [14]) highlights that although lacking specific histological features of an identifiable subset of muscle disease, the pathology underlying a non-specific myopathy may still progress. However, whether progression is due to inflammatory pathology not detected on B1 or due to disuse is not able to be ascertained from our data. The role of exercise (acknowledging the challenges of this in patients with muscle disease), to prevent irreversible muscle atrophy/damage in patients with histological finding of non-specific myopathy, is at present unclear.

One of the difficulties in assessing the clinical utility of repeat biopsies arises from the concern that obtaining a histopathological diagnosis may not necessarily affect decision making. By adopting a similar approach to Aburahma et al. in defining two outcomes of repeat biopsy [10], both of which result in a specific revision of the initial histopathological diagnosis, our findings support the role of repeat muscle biopsy in achieving a diagnosis and guiding management. A larger proportion (42%) of repeat muscle biopsies in our cohort met one of the two outcomes than previous reports which have ranged from 24% to 34% [10, 11]. Whether this attests to patient selection, biopsy technique, histological variation within a muscle independent of disease progression, or the more recent period of review reflecting advances in histopathological assessment of muscle is difficult to know.

In reviewing patients who have received serial muscle biopsies, we also had the opportunity to assess histopathological change over time, reflecting both disease evolution and immunomodulatory treatment effect. Interestingly, amongst patients not exposed to immunomodulatory treatment, repeat biopsy showed improvement of inflammation in a few cases. This does suggest that histology lacks sufficient sensitivity to be a monitoring tool. It is recognized that corticosteroid exposure does not appear to affect the inflammatory infiltrate of patients with PM [15] or DM [16], or MHC upregulation in adult and juvenile DM [17]. Whether non-corticosteroid immunomodulators attenuate histological changes, and whether the duration of immunomodulator exposure impacts the histopathology, is not known. These conclusions are restricted to a group of patients in whom either refractory/relapsing disease or diagnostic uncertainty triggered further biopsy.

Several limitations are acknowledged. The practices influencing a decision to perform muscle biopsy have changed over the last 24 years with increasing awareness of myositis-specific antibodies, access to genetic testing, and muscle MRI. Consequently, it stands to reason that the thresholds to consider a repeat muscle biopsy have also changed over the period of our analysis. Additionally, histopathological diagnosis can be affected by sampling error due to the patchy nature of muscle pathology [2] and needs to be considered within an appropriate clinical context; therefore, we are cautious to avoid extrapolating our data more broadly to patients with muscle conditions in general. Although Outcome 1 and 2 capture the discovery of a specific diagnosis on repeat biopsy, sometimes a non-specific or unchanged histological diagnosis may nonetheless be a clinically significant result through limiting the differential diagnosis or confirming a previously held suspicion. Heterogeneity is present not only in the patient cohort who underwent repeat biopsy but also in the requesting physicians. There is likely to be a broad range of practices about when to consider a repeat biopsy. Finally, we recognize that our cohort, by nature of having undergone a repeat biopsy, is highly selected and therefore not representative of all patients with neuromuscular pathology.

5 Conclusion

Repeat muscle biopsy has the highest yield when considering a diagnosis of IBM, and in patients whose initial biopsy shows non-specific myopathy or MNOS. Notably, with repeat biopsy, more than half of those with a non-specific myopathy gain diagnostic clarification, which impacts treatment. We show the potential for non-specific myopathy to progress to histological atrophy and highlight the importance of pursuing diagnosis and appropriate treatment in these patients. Our results affirm the role for an open surgical approach when repeating a muscle biopsy to increase the likelihood of obtaining a more specific diagnosis. As has previously been reported, inflammation on muscle biopsy evolves in an unpredictable fashion over time and with immunomodulation, suggesting heterogeneous pathophysiology and the limited utility of monitoring histology.

Author Contributions

Thomas Khoo: conceptualization, investigation, writing – original draft, methodology, writing – review and editing, formal analysis, data curation, project administration. Sarah Saxon: conceptualization, investigation, resources, data curation, formal analysis, visualization. Barbara Koszyca: conceptualization, investigation, data curation, resources, formal analysis, visualization. Bernice Gutschmidt: conceptualization, investigation, methodology, validation, formal analysis, data curation, resources. Vidya Limaye: conceptualization, investigation, writing – original draft, methodology, validation, writing – review and editing, formal analysis, data curation, supervision, resources.

Ethics Statement

We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Conflicts of Interest

The authors declare no conflicts of interest.