ERS clinical practice guidelines on treatment of sarcoidosis

Recommendation

For untreated patients with major involvement from pulmonary sarcoidosis believed to be at higher risk of future mortality or permanent disability from sarcoidosis, we recommend the introduction of glucocorticoid treatment to improve and/or preserve FVC and QoL. (Strong recommendation, low quality of evidence.)

Summary of evidence

The clinical outcomes identified by the panel included overall response, chest radiography and pulmonary function changes, and symptoms. Unfortunately, markers for increased morbidity or mortality were not specifically studied in the identified trials. Our systematic review identified 1747 potentially relevant articles; the full texts of 36 were reviewed and 19 were selected [28, 31, 75–91]. Many of our pre-specified outcomes were not evaluated in these trials.

The overall response to oral glucocorticoid treatment, based on clinical and radiological evaluation of two studies involving 134 patients [75, 76], found a larger proportion of patients experiencing clinical improvement (risk ratio 2.44, 95% CI 1.40–4.25) in short-term follow-up (3–6 months). There was also a trend towards fewer patients experiencing clinical deterioration (risk ratio 0.38, 95% CI 0.11–1.31), in the short term. In three placebo-controlled trials involving 340 patients [75, 77, 78], radiographic improvements favoured glucocorticoid treatment (risk ratio 1.35, 95% CI 1.11–1.64) with a lower prevalence of significant radiographic deterioration (risk ratio 0.39, 95% CI 0.18–0.87). Pulmonary function was not significantly impacted for the whole group [75, 77, 78], but there was a significant pulmonary function improvement for patients with initial lung involvement [77, 79]. Asymptomatic patients without radiographic improvement were randomly allocated to receive either glucocorticoids for at least 18 months or glucocorticoids only if clinically worsened. At 5 years the treated group had better functional outcome [80]. It should be stressed that these data may not apply to the subgroup of patients with higher risk disease. Interventions across the entire range of disease severity, including patients with limited or inactive disease, do not provide guidance in this important subgroup. This especially holds true for failure to demonstrate pulmonary function improvement in whole cohorts, including many patients with mild or intrinsically irreversible disease. Specifically, there is no existing controlled evaluation of glucocorticoid treatment efficacy in preventing pulmonary function decline in severe pulmonary disease.

Data from additional studies

Glucocorticoid treatment clearly has short-term efficacy by improving symptoms and chest radiography, and in achieving regression or prevention of progression in some cases. Currently, there is no suggestion that these effects are attenuated in higher risk disease. Based on two studies, these benefits appear to be short lived as they do not persist after discontinuation of glucocorticoids [76, 90]. The dose of glucocorticoids varied, but two studies found no additional benefit for treating pulmonary disease with more than 20 mg prednisone a day [81, 82]. It has been observed that at least half of patients started on glucocorticoids were still on treatment 2 years later [28, 31, 88]. None of the current studies or accumulated clinical experience specifically evaluated higher risk disease or whether stable disease with glucocorticoid treatment is likely to progress with the same glucocorticoid dosage. In summary, the data provide a basis for a likely long-term glucocorticoid treatment benefit in high-risk pulmonary sarcoidosis. To date, no data exist concerning mortality balance between benefits from long-term treatment and risks due to treatment-induced comorbidities. This underlines the importance of re-evaluating the need for glucocorticoid continuation in the longer term in chronic fibrotic pulmonary sarcoidosis unlikely to benefit from prolonged treatment.

Response to treatment for 3–6 months, if unsustained after treatment cessation [76, 83], provides a solid rationale to limit glucocorticoid use to patients with higher risk disease or unacceptable loss of QoL, or combined pulmonary and systemic symptoms.

In three double-blind, placebo-controlled randomised trials, the addition of inhaled glucocorticoid (versus placebo) to oral glucocorticoid did not provide significant benefits regarding symptoms or pulmonary function tests [84–86].

Justification of recommendations

Systemic glucocorticoid administration is associated with an overall response, as judged by a clinician, or based on clinical and radiological evaluation. It is also associated with radiological improvement. The strong recommendation for glucocorticoid use in symptomatic pulmonary patients at risk for mortality is based on data summarised in supplement S2 in the supplementary material and includes several randomised trials [75, 76, 78–80, 90–92]. This strong recommendation was based on the committee's consensus concerning a serious situation warranting treatment.

For patients with worsening QoL from pulmonary disease, we recommend shared decision making between physicians and patients with a consideration of initial low- to medium-dose glucocorticoid treatment (5–10 mg a day) [4], and with the dose and duration of maintenance treatment based on the efficacy/side-effects balance.

For patients not felt to be at risk for morbidity or mortality, or having no significant impairment of QoL, the Task Force usually offers no glucocorticoid treatment because of the high prevalence of adverse events. Figure 1 summarises this approach.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

Approach for pulmonary sarcoidosis. Use of rituximab, JAK inhibitor and repository corticotropin injection (RCI) should be on a case-by-case basis. This figure is a combination of the recommendations made in this guideline and a description of Task Force members’ current practice in situations where there was not enough evidence to warrant a recommendation or for questions for which a systematic review of the literature was not undertaken. Note that the information depicted as current practice (in white boxes) is not intended as a recommendation for clinical practice. ^#: assess need for treatment based on low risk, intermediate risk but impaired quality of life or high risk as discussed in text. GC: glucocorticoid.

Future research

There is an urgent need for accurate risk stratification in pulmonary sarcoidosis. Unmet needs include optimal pulmonary function thresholds, integrated with disease duration and risk assessment for progression in higher risk disease. It is uncertain when higher risk disease is best managed with glucocorticoid monotherapy as opposed to combination therapy with second- or third-line agents. The role of PET in rationalising long-term treatment following initial stabilisation of irreversible disease requires exploration in large cohorts.

A database is needed to quantify the therapeutic efficacy of glucocorticoids in patients with unacceptable loss of QoL, to explore the efficacy and adverse effects with the use of low-dose glucocorticoid treatment, and to evaluate the optimal dose and duration driven by patient choice.

Another area which needs to be better studied includes how high the initial glucocorticoid dosage should be, how long to stay on that dose and how to taper.

Recommendations

Recommendation 1: For patients with symptomatic pulmonary sarcoidosis believed to be at higher risk of future mortality or permanent disability from sarcoidosis who have been treated with glucocorticoids and have continued disease or unacceptable side-effects from glucocorticoids, we suggest the addition of methotrexate to improve and/or preserve FVC and QoL. (Conditional recommendation, very low quality of evidence.)

Recommendation 2: For patients with symptomatic pulmonary sarcoidosis believed to be at higher risk of future mortality or permanent disability from sarcoidosis who have been treated with glucocorticoids or other immunosuppressive agents and have continued disease, we suggest the addition of infliximab to improve and/or preserve FVC and QoL. (Conditional recommendation, low quality of evidence.)

Summary of evidence

Studied populations include patients with chronic symptomatic pulmonary sarcoidosis treated with glucocorticoids and/or other immunosuppressive agents. The systematic literature review identified 1319 potentially relevant articles; the full texts of 41 were reviewed and six were selected [93–98]. We identified six drugs with adequate reports: infliximab, golimumab, ustekinumab, pentoxifylline, cyclosporine A and methotrexate. As displayed in the evidence to decision table (supplement S2 in the supplementary material), most of our pre-selected outcomes were not evaluated in clinical studies or trials. Some randomised controlled interventions were studied in patients receiving glucocorticoid. Infliximab, compared to prednisone, significantly improved FVC, the primary end-point in two phase III randomised trials for the treatment of chronic respiratory symptoms. However, absolute FVC changes were small. Secondary end-points included chest imaging and QoL assessments [94, 96].

In one randomised, double-blind, placebo-controlled trial, methotrexate did not demonstrate significant FVC improvement, although allowing a significant prednisone reduction with lower weight gain in the second 6 months [93]. Other open-label prospective and retrospective trials have found methotrexate to be steroid sparing and associated with improved lung function [45, 99, 100].

No recommendation could be made for cyclosporine A, golimumab or ustekinumab as randomised trials showed no benefit over placebo [95, 98]. These drugs should be considered on a case-by-case basis.

Data from additional studies

Azathioprine is as effective as methotrexate in pulmonary sarcoidosis [45, 101]. Leflunomide and mycophenolate mofetil are also effective [44, 46, 102]. In a randomised trial, chloroquine was mildly beneficial in pulmonary sarcoidosis [103]. In a retrospective study from one centre, chloroquine was less effective than in skin sarcoidosis [104]. Adalimumab was found effective for pulmonary disease in a prospective, open-label trial [47] and a small retrospective series [105].

Some studies support the use of rituximab [106]. The CLEAR (Concomitant Levofloxacin, Ethambutol, Azithromycin and Rifampin) regimen was found effective in a small uncontrolled observational study [107], but a recently reported double-blind, placebo-controlled trial found no difference in response rate compared to placebo [108]. The committee did not feel that current data supported a treatment recommendation for CLEAR. Repository corticotropin injection has been found to be steroid sparing in two retrospective studies [109, 110] and one prospective study [48]. However, the drug is currently quite expensive and the mechanism of action remains unclear [111]. There is a reported response to a JAK inhibitor and benefits with anti-interleukin-6 therapy in small retrospective series [112, 113]. These agents are considered by the Task Force members on a case-by-case basis when other therapies are ineffective or not tolerated.

Justification of recommendation

The evidence base of the conditional recommendation for methotrexate in symptomatic pulmonary patients at risk for mortality is summarised in supplement S2 in the supplementary material and includes a randomised trial [93]. The conditional recommendation for infliximab in symptomatic chronic pulmonary sarcoidosis not responding to other immunosuppressives, including glucocorticoids, is based on two trials summarised in supplement S2 in the supplementary material [94, 96]. The committee could not make recommendations on other drugs. Data supporting the use of some drugs is provided in the evidence to decision table in supplement S2 in the supplementary material. Figure 1 summarises the approach used by most members of committee.

Future research

Additional studies are needed to evaluate the efficacy, safety and cost efficiency of rituximab, repository corticotropin injection, anti-tumour necrosis factor (TNF) biosimilars and other immunosuppressive agents. Also, the role of antifibrotic agents such as nintedanib and pirfenidone needs to be further studied [114]. Newer end-points, including change in PET and QoL, need to be validated.

Recommendation

For patients with cutaneous sarcoidosis and cosmetically important active skin lesions which cannot be controlled by local treatment, we suggest oral glucocorticoids be considered to reduce skin lesions. (Conditional recommendation, very low quality of evidence.)

Summary of evidence

This question was originally framed to study patients with extrapulmonary sarcoidosis treated with glucocorticoids versus no treatment. It was narrowed to study patients with cutaneous sarcoidosis when the systematic literature review revealed that this population was the focus of the preponderance of studies in this area. Clinical outcomes identified by the panel as being important included clinical remission and remission of lupus pernio.

Our systematic literature review identified 1032 potentially relevant articles; the full texts of 33 were reviewed and seven were selected [120, 125–130]. As seen in the evidence to decision table in supplement S2 in the supplementary material, most of our pre-selected outcomes were not evaluated in the trials that we studied. The two outcomes assessed were clinical remission and remission of lupus pernio, as reported by the authors.

There were no randomised trials in this area. We selected six retrospective observational cohort studies on skin sarcoidosis with different types of lesions and localisations, all of which studied at least 20 patients [125–130]. Treatment with systemic glucocorticoids was associated with improvement or remission in up to two-thirds of patients. Often, the desired effects were limited to the duration of treatment and recurrences were not uncommon upon glucocorticoid tapering, requiring additional immunosuppressive therapy. For patients with lupus pernio, a retrospective study on 54 patients showed that only 20% of patients receiving systemic glucocorticoid alone achieved complete or near-complete resolution and 50% having some improvement but requiring an average daily prednisone dose of 16 mg [120]. This study employed evaluating photographs of the lesions before and after treatment, but the assessment was retrospective and photographs were obtained at various times during therapy.

Data from additional studies

Topical glucocorticoids are generally considered to be beneficial for limited skin lesions of mild or moderate extension. However, evidence of their efficacy is scarce. In a study of 20 patients who received topical treatment including intralesional administration, only five had complete resolution and the rest had partial resolution [125]. Clobetasol or halobetasol propionate have been used especially for limited and discrete papules and plaque [131, 132]. Intralesional injections of triamcinolone acetonide may be more effective than topical preparations [133]. Topical or intralesional glucocorticoids are impractical for cases with widespread lesions [134].

Justification of recommendation

The conditional recommendation for glucocorticoids for cosmetically important skin lesions is based on the few retrospective studies which reported resolution of lesions. The short-term response was commonly seen. There was insufficient evidence to make a recommendation regarding topical glucocorticoids. While physicians are comfortable with using glucocorticoids, the risk of long-term adverse effects must always be considered.

Implementation consideration

While oral glucocorticoids were effective, prolonged use is associated with substantial side-effects. Use of steroid-sparing alternatives should be considered whenever possible, especially for chronic lesions such as lupus pernio.

Future research

With the advent of new technologies to assess skin response, the value of topical and systemic glucocorticoids should be re-evaluated. Among the new tests are standardised skin scoring techniques [56, 57]. The role of high-frequency ultrasound to assess skin lesions needs further evaluation [135].

Recommendation

For patients with cutaneous sarcoidosis who have been treated with glucocorticoids and/or other immunosuppressive agents and have continued cosmetically important active skin disease, we suggest the addition of infliximab compared to no additional treatment to reduce skin lesions. (Conditional recommendation, low quality of evidence.)

Summary of evidence

This question was originally framed to study patients with extrapulmonary sarcoidosis treated with immunosuppressive treatments compared to those receiving glucocorticoids. It was narrowed to study patients with cutaneous sarcoidosis when the systematic literature review revealed that this population was the focus of the preponderance of studies in this area. Clinical outcomes identified by the panel as being important included a validated metric for assessing cutaneous lesions (SASI score [56, 57]) and QoL metrics (Short Form-36 (SF-36) Physical Component Summary and SF-36 [136]).

Our systematic literature review identified 980 potentially relevant articles. The full texts of 91 articles were reviewed. We identified five prospective controlled studies of patients with cutaneous sarcoidosis randomised to either an immunosuppressive agent or continuing glucocorticoid that had quantitative data amenable to extraction [95, 124, 137–139]. We identified two prospective, randomised controlled studies that compared the use of infliximab to glucocorticoid to treat cutaneous sarcoidosis and provided data concerning our selected outcomes [124, 138]. Baughman et al. [124] demonstrated a statistically significant improvement in the SASI desquamation index in patients treated with infliximab compared to glucocorticoid alone. In an additional study, an extrapulmonary organ severity tool (ePOST) was used to assess individual organ involvement [138]. The ePOST tool was useful as a broad assessment of each organ, but it was not specific for skin involvement.

Data from additional studies

Two randomised trials using drugs targeted against TNF other than infliximab failed to show benefit for treating cutaneous sarcoidosis. One was for golimumab [95] and the other was for thalidomide [137]. The latter study used different end-points than a previous positive open-label trial of thalidomide for cutaneous sarcoidosis [140]. Adalimumab (also a monoclonal antibody against TNF) has also been studied in one double-blind, placebo-controlled trial and was found to be more effective than placebo for chronic cutaneous sarcoidosis [139]. This study was not abstracted for analysis because only qualitative data were available. Future studies are needed to explore the clinical benefit of adalimumab.

Other treatments have been used for cutaneous sarcoidosis that have not been studied in prospective, randomised controlled studies. There has been an open-label prospective trial of treatment for sarcoidosis using chloroquine [141]. The positive response to chloroquine has been confirmed by other case series, many of which included hydroxychloroquine instead of chloroquine [104, 125, 142]. Methotrexate has been reported as effective in treating cutaneous disease in several series for both adults and children [99, 143–145]. There has been an open-label prospective trial of treatment for sarcoidosis with apremilast [122]. The positive response to apremilast has not been confirmed by either case series or another clinical trial. There have been no clinical series reporting on the use of azathioprine, leflunomide or mycophenolate mofetil specifically for cutaneous sarcoidosis. These drugs have been reported as useful for chronic sarcoidosis [44–46, 102]. However, none of these drugs has been specifically studied in cutaneous sarcoidosis, so we are unable to make recommendations regarding their use.

We identified one additional study that examined the CLEAR regimen instead of an immunosuppressive agent and compared this to placebo to treat patients with cutaneous sarcoidosis [123]. Both an intention-to-treat and a per-protocol analysis demonstrated a statistically significant improvement in the SASI score with CLEAR treatment. The CLEAR trial was a single-masked study performed at one centre and has not been confirmed. However, a subsequent larger double-blind, placebo-controlled trial of CLEAR for pulmonary disease found no evidence of effectiveness of this regimen [108]. The committee did not feel that current data supported a treatment recommendation.

Justification of the recommendation

Two small, prospective, randomised controlled studies demonstrate improvement in sarcoidosis cutaneous lesions as assessed by the SASI score with treatment by infliximab compared to continued glucocorticoid and other immunosuppressants alone in patients with cutaneous sarcoidosis [94, 124]. Infliximab is an immunomodulatory agent with a risk of adverse effects including increased susceptibility to infection, although adverse events were low in the analysed studies. The balance of effects would lead most patients to favour the use of infliximab. The stepwise approach to the management of cosmetically important cutaneous sarcoidosis is shown in figure 2.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2

Stepwise approach to the management of cosmetically important cutaneous sarcoidosis. Use of apremilast and tofacitinib should be on a case-by-case basis. This figure is a combination of the recommendations made in this guideline and a description of Task Force members’ current practice in situations where there was not enough evidence to warrant a recommendation or for questions for which a systematic review of the literature was not undertaken. Note that the information depicted as current practice (in white boxes) is not intended as a recommendation for clinical practice. ^#: assess need for treatment as discussed in text. GC: glucocorticoid.

Implementation considerations

Barriers to use of infliximab include the expense of treatment, the availability of facilities for parenteral administration of the agent and the potential of adverse effects. Some patients might wish to avoid agents that require parenteral administration.

Future research

The skin is an easy organ to assess, resample and biopsy. This makes it a useful target for evaluating new therapies in sarcoidosis. It is important to show whether changes in the skin reflect other organ involvement.

Recommendation

For patients with evidence of functional cardiac abnormalities, including heart block, dysrhythmias or cardiomyopathy, we recommend the use of glucocorticoids (with or without other immunosuppressives). (Strong recommendation, very low quality of evidence.)

Summary of evidence

For this PICO, the clinical outcomes included: improvement, worsening/progression (defined by several findings AND clinical judgement), changes in cardiac PET, changes in cardiac MRI, arrythmias, QoL and toxicity [155, 163–165]. Our systematic literature review identified 996 potentially relevant articles; the full texts of 33 were reviewed and 17 were selected [67, 150, 153, 155, 157, 158, 166–176]. The data included retrospective studies specifically examining the effect of glucocorticoid treatment versus no treatment and association studies that included glucocorticoid therapy as a covariate predictor of various cardiac outcomes. No study that specifically assessed the effects of glucocorticoid therapy enrolled patients prospectively or systematically with sufficient rigour to directly compare the outcomes; all studies were subject to substantial risk-of-channelling bias or other unmeasured confounders. However, the available data suggest that the risks of important composite cardiac end-points were reduced, with hazard ratios ranging from 0.33 to 0.78. Many of the end-point events were driven by appropriate defibrillator or antiarrhythmic therapies, which were inferred (but not proven) to be equivalent to the prevention of sudden cardiac death. Nonetheless, the bulk of the studies evaluated outcomes deemed likely to be of critical importance to affected patients.

Data from additional studies

Heart block is often an early sign of cardiac involvement and it may be the manifestation with the best chance of responding to glucocorticoids [167, 177]. The optimal dose and duration of immunosuppressive therapy are unknown. A retrospective analysis suggested that prednisolone doses higher than 0.5 mg·kg⁻¹ were no more effective than a starting dose of 0.5 mg·kg⁻¹ [178]. It is likewise unclear whether pulse intravenous methylprednisolone is useful and for whom should it be considered [179]. Some data suggest that earlier initiation of glucocorticoids confers better cardiac outcomes [165]. Similarly, one retrospective case–control study found that withdrawal of glucocorticoids after initiation of treatment, regardless of clinical improvement, was associated with worse outcomes [180].

Glucocorticoids may lead to significant morbidity [181]; therefore, early initiation of steroid-sparing medications should be considered [4]. However, for cardiac sarcoidosis, the evidence to support steroid-sparing medications is poor and subject to all the aforementioned biases. The most commonly described steroid-sparing agents were methotrexate, azathioprine, mycophenolate mofetil, leflunomide and cyclophosphamide [154, 155, 182]. In most of the studies, the patients treated with steroid-sparing agents had no better outcomes that those treated with glucocorticoid monotherapy, but a single-centre retrospective study comparing the addition of methotrexate to prednisone versus prednisone alone suggested improved ejection fraction and brain natriuretic peptide after 5 years of treatment [164]. Anti-TNF antibodies may be useful for refractory disease [183, 184].

Justification of the recommendation

The level of evidence to support treatment approaches for cardiac sarcoidosis was very low, with multiple potential confounders and biases inherent in the available studies [149, 185]. Much of the data supporting the use of glucocorticoids is indirect, originating in association studies where glucocorticoid treatment is a covariate among other outcome predictors [185]. There is likewise minimal description in the available studies of the indications for glucocorticoid treatment or the characteristics of the treated versus untreated patients. The risk of death from cardiac sarcoidosis is high, especially for those with reduced left ventricular function [153]. Since glucocorticoid treatment has been associated with improvement in left ventricular ejection [155, 165], the Task Force members concluded that the danger associated with cardiac sarcoidosis favoured glucocorticoid treatment for clinically relevant cardiac sarcoidosis [5, 186]. There was insufficient evidence to make a recommendation regarding other immunosuppressants, but the Task Force members still consider such treatment to minimise the toxicity of glucocorticoids. Figure 3 summarises the approach used by most Task Force members.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3

Approach to cardiac sarcoidosis. This figure is a combination of the recommendations made in this guideline and a description of Task Force members’ current practice in situations where there was not enough evidence to warrant a recommendation or for questions for which a systematic review of the literature was not undertaken. Note that the information depicted as current practice (in white boxes) is not intended as a recommendation for clinical practice. ^#: use of implanted cardioverter defibrillator recommendation adapted from the Heart Rhythm Society [160, 162]; ^¶: clinically relevant cardiac sarcoidosis is defined as rhythm disturbances, heart failure or high risk for sudden cardiac death; ⁺: infliximab and adalimumab are usually used in combination with second-line agents. GC: glucocorticoids.

Future research

An area of current uncertainty is the management of asymptomatic patients with concerning imaging features, such as late gadolinium enhancement, fluorodeoxyglucose uptake, T2 prolongation or impaired global longitudinal strain, even when cardiac function is preserved and electrical abnormalities are absent [163, 187].

Other issues include the optimal dose of glucocorticoid, duration of treatment and role of steroid-sparing medications. There is an urgent need to develop and validate reliable biomarkers and imaging features for the assessment of treatment response.

Recommendations

Recommendation 1: For patients with clinically significant neurosarcoidosis, we recommend treatment with glucocorticoids. (Strong recommendation, very low quality of evidence.)

Recommendation 2: For patients with neurosarcoidosis that have been treated with glucocorticoids and have continued disease, we suggest the addition of methotrexate. (Conditional recommendation, very low quality of evidence.)

Recommendation 3: For patients with neurosarcoidosis that have been treated with glucocorticoids and a second-line agent (methotrexate, azathioprine, mycophenolate mofetil) and have continued disease, we suggest the addition of infliximab. (Conditional recommendation, very low quality of evidence.)

Summary of evidence

The clinical outcomes that were evaluated were: improvement, worsening/progression (defined by several findings AND clinical judgement), QoL and toxicity. Our systematic literature review identified 1305 potentially relevant articles; the full texts of 56 were reviewed and four were selected [35, 191, 194, 195].

One retrospective analysis of 234 neurosarcoidosis patients [191] found that although treatment with glucocorticoid alone significantly lowered the overall relapse rate of sarcoidosis compared to no treatment (hazard ratio 0.59, 95% CI 0.39–0.90; p=0.01), the specific rate of neurosarcoidosis relapse was not significantly affected (hazard ratio 0.68, 95% CI 0.38–1.23; p=0.2). Additional drugs besides glucocorticoids were found to significantly lower the relapse rate of neurosarcoidosis in this cohort (see the following data section) and most of these drugs were used in combination with glucocorticoids; this suggests glucocorticoids may have contributed to protecting against neurosarcoidosis relapse in these cases. In a meta-analysis of 1088 neurosarcoidosis patients [194], glucocorticoids were initiated as first-line therapy in 434 out of 539 (81%) treated patients and a favourable outcome was reported in 161 out of 227 (71%, 95% CI 65–77%) patients who only received glucocorticoid. We believe that these data, although limited, support the use of glucocorticoids as first-line therapy for neurosarcoidosis.

Joubert et al. [191] demonstrated that infliximab statistically significantly lowered the rate of overall sarcoidosis relapse (hazard ratio 0.31, 95% CI 0.11–0.82; p=0.02), but failed to demonstrate a statistically significant lower relapse rate of neurosarcoidosis (hazard ratio 0.16, 95% CI 0.02–1.24; p>0.05). A retrospective report demonstrated good neuroimaging and functional outcomes in 66 neurosarcoidosis patients treated with infliximab-containing regimens [35].

Data from additional studies

Reports of treatment of neurosarcoidosis consist of the second-line agents methotrexate, azathioprine and mycophenolate mofetil as well as antimalarial drugs and cyclosporine A. These drugs are usually added to glucocorticoid treatment when glucocorticoids are ineffective or a relapse occurs after tapering. These drugs may be used concomitantly with glucocorticoids as part of the initial treatment of neurosarcoidosis. The evidence for these agents is also sparse, with the possible exception of methotrexate [196]. An analysis from one institution found a statistically significant reduction in the relapse rate of neurosarcoidosis with methotrexate (hazard ratio 0.47, 95% CI 0.25–0.87; p=0.02) and hydroxychloroquine (hazard ratio 0.37, 95% CI 0.15–0.92; p=0.03), but not with azathioprine (hazard ratio 1.88, 95% CI 0.69–5.14; p=0.22) or mycophenolate mofetil (hazard ratio 0.58, 95% CI 0.25–1.34; p=0.20) [191]. In the previously described meta-analysis [194], treatment with methotrexate, azathioprine and hydroxychloroquine was initiated in 144 of the 539 (27%) patients who were treated for neurosarcoidosis. A favourable outcome was observed in 47 of the 85 (55%, 95% CI 45–66%) patients who received these agents and were not switched to third-line therapy. A retrospective analysis was performed concerning 40 neurosarcoidosis patients who received either methotrexate (n=32) and/or mycophenolate mofetil (n=14) as part of their treatment regimen [195]. Those who received methotrexate had a significantly lower yearly relapse rate than those who received mycophenolate mofetil (0.2 versus 0.6 relapses per year; p=0.058) and the median time to relapse was also longer in the methotrexate group (28 versus 11 months; p=0.049). To summarise the available data concerning the use of non-biologic agents for the treatment of neurosarcoidosis, the limited data support the use of methotrexate. Although the evidence for the other agents is minimal, there is inadequate evidence to state that these agents are ineffective for neurosarcoidosis. After methotrexate, we would consider azathioprine, mycophenolate mofetil or hydroxychloroquine. Although chloroquine and cyclosporine A could also be considered as potential second-line agents for neurosarcoidosis, their side-effect profile suggests that other non-biologic agents should be preferred. We are only aware of two case reports suggesting that adalimumab is beneficial for the treatment of neurosarcoidosis [197, 198]. There is low quality of evidence supporting cyclophosphamide for the treatment of neurosarcoidosis. In one study [191], intravenous cyclophosphamide statistically significantly lowered the rate of relapse of neurosarcoidosis compared to untreated patients (hazard ratio 0.26, 95% CI 0.11–0.59; p=0.001). In addition, in a retrospective series [196, 199], cyclophosphamide was found to be beneficial for neurosarcoidosis that was refractory to glucocorticoids and methotrexate. Despite the potential efficacy of cyclophosphamide for the treatment of neurosarcoidosis, we believe that infliximab and even adalimumab are more preferred based on the side-effect profiles of these agents.

Justification of recommendation

The strong recommendation for glucocorticoids for clinically significant neurosarcoidosis is based on very low quality of evidence; the committee felt the high risk for significant irreversible neurologic loss warranted the strong recommendation. The conditional recommendation for infliximab was based on two retrospective studies [35, 200] and other studies, as summarised in supplement S2 in the supplementary material.

Clinical evidence concerning the treatment of neurosarcoidosis is meagre due to the absence of any randomised controlled trial and to the wide variety of outcomes evaluated in retrospective studies (neuroimaging, remission/relapse, functional status, mortality) which evaluated different drugs. In addition, because drug trials for neurosarcoidosis have not rigorously compared specific agents against other ones, our recommendations concerning the stepwise approach to the treatment of neurosarcoidosis are based not only on efficacy data but also drug cost, side-effect profile and ease of use. Figure 4 shows the committees usual approach to treating neurosarcoidosis.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4

Approach to neurologic sarcoidosis. This figure is a combination of the recommendations made in this guideline, and a description of Task Force members’ current practice in situations where there was not enough evidence to warrant a recommendation or for questions for which a systematic review of the literature was not undertaken. Note that the information depicted as current practice (in white boxes) is not intended as a recommendation for clinical practice. ^#: infliximab and adalimumab are usually used in combination with second-line agents. GC: glucocorticoids.

Future research

Studies confirming the effectiveness of infliximab for neurosarcoidosis need to be performed. Studies examining whether high-dose glucocorticoids are required with infliximab as initial treatment for advanced neurosarcoidosis may help reduce the burden of glucocorticoid toxicity. These studies would require standardised outcome measures. Given the relative rarity of neurosarcoidosis, multicentre studies will most likely be required. In addition, neurosarcoidosis may not be amenable to uniform treatment decisions but may require different treatments depending both on the localisation and the severity of affection (central, peripheral or spine).

Recommendations

Recommendation 1: In patients with sarcoidosis who have troublesome fatigue, we suggest a pulmonary rehabilitation programme and/or inspiratory muscle strength training for 6–12 weeks to improve fatigue. (Conditional recommendation, low quality of evidence.)

Recommendation 2: In patients with sarcoidosis who have troublesome fatigue that is not related to disease activity, and after consideration of a pulmonary exercise or rehabilitation programme, we suggest the use of d-methylphenidate or armodafinil for 8 weeks to test its effect on fatigue and tolerability. (Conditional recommendation, low quality of evidence.)

Summary of evidence

Our systematic literature review of articles regarding fatigue and sarcoidosis identified 165 potentially relevant articles; the full texts of 27 were reviewed and five were selected [205–209]. One of these was of an experimental intervention not available at this time (cibinetide) [209]. The remaining four articles were reviewed.

Two of the interventions involved randomised controlled trials with physical therapist interventions. Inspiratory muscle training for 6 weeks has been studied, which led to significant improvement of the 6-min walk test (6MWT), Borg dyspnoea scale, maximal inspiratory and expiratory pressure, and Fatigue Severity Scale in the treatment group [208]. A second randomised controlled trial has tested the effect of a structured exercise programme for 12 weeks [207]. Significant effects were found on the following outcomes: 6MWT, Borg dyspnoea scale, modified Medical Research Council dyspnoea scale, maximal inspiratory force, leg strength, partial arterial oxygen pressure, and Fatigue Severity Scale and St George's Respiratory Questionnaire.

Pharmacologic interventions with neurostimulants have also been evaluated by two randomised controlled trials. Dexmethylphenidate hydrochloride (d-methylphenidate (d-MPH)) was given to 10 patients with median (range) Functional Assessment of Chronic Illness Treatment-Fatigue (FACIT-F) score of 16 (4–37) and Fatigue Assessment Scale (FAS) of 38 (22–44) in a randomised crossover trial [205]. The improvement in fatigue at 8 weeks for the d-MPH group was 36%, similar to the improvement seen in patients with cancer chemotherapy-related fatigue [210]. In that study, no difference in toxicity was noted between drug and placebo. The other randomised controlled trial investigated armodafinil 150 mg daily for 4 weeks, then 250 mg daily for 4 weeks [206]. This resulted in an improvement in fatigue as measured by the FAS and FACIT-F scores. Only 15 patients were studied. One patient withdrew because of anxiety. The adverse effects of methylphenidate and armodafinil are known from other patient populations, and include addiction, insomnia, anxiety and tachycardia [211].

Data from additional studies

Other observational studies have shown positive effects of exercise training or rehabilitation programmes on SAF and other parameters associated with reduced QoL [212–214]. One study demonstrated improvement in fatigue as well as 6MWD for those participating in pulmonary rehabilitation [215]. A recent randomised trial, published since our systematic literature review, found that rehabilitation improved fatigue [216]. This regimen was comparable to other pharmacologic interventions [217]. A recent randomised controlled trial also showed that the use of low-dose glucocorticoids has been shown to alleviate SAF, especially in the context of ongoing inflammation [218], but the committee felt there was insufficient evidence to make a recommendation regarding low-dose glucocorticoids.

Justification of recommendation

The conditional recommendations for the treatment for SAF were each supported by one prospective trial. In the cases of physical treatment intervention, one study used a sham procedure for control and the other compared patients who chose not to participate in structured training. The pharmacologic interventions were both studied in a double-blind, placebo-controlled, crossover design. However, only a limited number of subjects were studied. The approach to the evaluation and management of SAF is shown in figure 5.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5

Approach to the evaluation and management of sarcoidosis-associated fatigue. The use of low-dose corticosteroids with or without methotrexate should be considered on a case-by-case basis. This figure is a combination of the recommendations made in this guideline, and a description of Task Force members’ current practice in situations where there was not enough evidence to warrant a recommendation or for questions for which a systematic review of the literature was not undertaken. Note that the information depicted as current practice (in white boxes) is not intended as a recommendation for clinical practice. ^#: other causes of fatigue include diabetes mellitus, thyroid dysfunction, neuroendocrine disorders, sleep apnoea, small-fibre neuropathy, vitamin D deficiency with low 1,25-dihydroxycholecalciferol, congestive heart failure and neurologic disease.

Future research

Further research is needed to confirm the effects of inspiratory muscle training, which have been noted in a single study, and to review the impact of the recommendation regarding physical training upon costs, resources and healthcare equity. The long-term effects should also be explored, especially how improvement can best be maintained after the end of training or a systematic rehabilitation programme.

Further research is needed to confirm the effects and toxicity of d-MPH and armodafinil, which have been noted in two single-centre studies, and to review the impact of the recommendation upon costs, resources and healthcare equity. The effects of long-term use of d-MPH and armodafinil should be explored.

Summary of evidence

Our systematic literature review identified 427 potentially relevant articles; the full texts of nine were reviewed and four were selected. Three of these involved cibinetide [209, 226, 227], an erythropoietin analogue, which is currently not available for clinical use. The other was a large retrospective review from one centre evaluating IVIg and/or anti-TNF monoclonal antibody treatment [17]. There are no validated, widely available end-points for evaluating the effect of SFN treatment in patients with sarcoidosis [16, 219, 221]. The clinical outcomes that were evaluated in this analysis were: measures of pain, measures of SFN (QSART), skin biopsies, SFN scale, cognitive scale and confocal microscopy. We were not able to identify sufficient treatment evidence to warrant a recommendation for any commercially available agent.

Data from additional studies

Treatment of SFN depends on the underlying disease, if identified. Symptoms are often disabling and difficult to alleviate, even when the cause is identified and adequately treated, leading to high morbidity and decreased QoL [219]. Usually, only symptomatic relief of complaints can be achieved. Guidelines for neuropathic pain have been adapted from the treatment regimens developed for other causes of SFN-related pain [219, 222]. There is no consensus regarding evaluating outcome for response to specific therapy for SFN. To date, studies have evaluated improvement in the autonomic symptoms, fibre neuropathy symptoms and the related pain, and the number of small fibres in the cornea [209, 226, 227]. However, these have not been routinely applied and were not employed in retrospective reports [17, 228, 229].

A large observational study found that that 75% of patients derived symptomatic benefit from a dosing regimen of IVIg either alone or in conjunction with anti-TNF monoclonal antibody therapy. The dosing regimen was like that described for chronic inflammatory demyelinating polyneuropathy [17]. A total of 79 patients were treated with IVIg alone or with anti-TNF monoclonal antibodies and were evaluated [17, 228]. The data are limited by the absence of a defined standard for assessing treatment response, patient selection bias, differences in concomitant treatment regimens and lack of a placebo group. Thus, conclusions regarding the usefulness of IVIg are currently preliminary. Nonetheless, a significant subset of patients are observed to experience moderate to dramatic improvements in symptoms and functionality within several months of initiating treatment [228]. The putative mechanism for effectiveness of IVIg is unclear, but may relate to immunomodulatory effects [230].

TNF may be a proximate trigger for central and peripheral inflammatory cascades that are postulated to cause neuropathy, as well as sarcoidosis itself [231]. The monoclonal anti-TNF antagonists infliximab and adalimumab have been assessed in two retrospective cohorts totalling 115 patients [17, 232]. These reports suggested that SFN-associated symptoms may respond to TNF inhibition, although the magnitude of the effect is difficult to ascertain from the available data. The GG promoter variant, associated with less exuberant TNF transcription, was also associated with better outcomes in treated patients [232].

Cibinetide, previously known as ARA-290, is an innate repair receptor agonist that has anti-inflammatory and neuroprotective properties [233–235]. Cibinetide is not yet approved for any indication, so it is not the subject of a formal recommendation in these guidelines. More importantly, it is not currently commercially available. However, cibinetide is the most extensively studied and best-validated treatment to date for sarcoidosis-associated SFN. In three randomised, placebo-controlled, double-blind studies, it has been shown to reduce symptom scores and improve markers of corneal nerve fibre health over short time-frames [209, 226, 227]. Interestingly, these neuropathic benefits correlated with increases in the 6MWD, underscoring the important functional consequences of SFN [209, 227, 234].

Justification

There were no studies with sufficient results to support any specific recommendations for SFN due to sarcoidosis. However, we have presented the current practice of managing SFN, summarised in figure 6.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6

An approach to small-fibre neuropathy symptoms used by Task Force members. The use of intravenous immunoglobulin (IVIg) or anti-tumour necrosis factor (TNF) antagonists should be considered on a case-by-case basis. This figure is a combination of the recommendations made in this guideline and a description of Task Force members’ current practice in situations where there was not enough evidence to warrant a recommendation or for questions for which a systematic review of the literature was not undertaken. Note that the information depicted as current practice (in white boxes) is not intended as a recommendation for clinical practice. GABA: γ-aminobutyric acid.

Future research

Safety and clinical effectiveness of cibinetide, IVIg, anti-TNF antibodies and other interventions for patients with sarcoidosis and SFN needs to be investigated. Development and clinical validation of accurate biomarkers and/or clinical scores to assess treatment response should be developed.