Cyclophosphamide for multiple sclerosis

Abstract

Background

Multiple sclerosis is a presumed cell‐mediated autoimmune disease of the central nervous system. Cyclophosphamide (CFX) is a cytotoxic and immunosuppressive agent, used in systemic autoimmune diseases. Controversial results have been reported on its efficacy in MS. We conducted a systematic review of all relevant trials, evaluating the efficacy of CFX in patients with progressive MS.

Objectives

The main objective was to determine whether CFX slows the progression of MS.

Search methods

We searched the Cochrane MS Group Trials Register (June 2006), Cochrane Central Register of Controlled Trials (The Cochrane Library Issue 3, 2006), MEDLINE (PubMed)(January 1966 to June 2006), EMBASE (January 1988 to June 2006) and reference lists of articles. We also contacted researchers in the field.

Selection criteria

Randomised controlled trials (RCTs) evaluating the clinical effect of CFX treatment in patients affected by clinically definite progressive MS. 
 CFX had to be administered alone or in combination with adrenocorticotropic hormone (ACTH) or steroids. The comparison group had to be placebo or no treatment or the same co‐intervention (ACTH or steroids)

Data collection and analysis

Two reviewers independently decided the eligibility of the study, assessed the trial quality and extracted data. We also contacted study authors for original data.

Main results

Of the 461 identified references, we initially selected 70: only four RCTs were included for the final analysis. Intensive immunosuppression with CFX (alone or associated with ACTH or prednisone) in patients with progressive MS compared to placebo or no treatment (152 participants) did not prevent the long‐term (12, 18, 24 months) clinical disability progression as defined as evolution to a next step of Expanded Disability Status Scale (EDSS) score. However, the mean change in disability (final disability subtracted from the baseline) significantly favoured the treated group at 12 (effect size ‐ 0.21, 95% confidence interval ‐ 0.25 to ‐0.17) and 18 months (‐ 0.19, 95% confidence interval ‐ 0.24 to ‐ 0.14) but favoured the control group at 24 months (0.14, CI 0.07 to 0.21). We were unable to verify the efficacy of other schedules. Five patients died; sepsis and amenorrhea frequently occurred in treated patients (descriptive analysis).

Authors' conclusions

We were unable to achieve all of the objectives specified for the review. This review shows that the overall effect of CFX (administered as intensive schedule) in the treatment of progressive MS does not support its use in clinical practice.

Plain language summary

The effect of the immunosuppressive drug cyclophosphamide in people with multiple sclerosis

MS is a chronic disease of the nervous system affecting young and middle‐aged adults. MS is supposed to be related to the immune system. CFX is an immunosuppressive drug used for various autoimmune diseases. As its use for MS is controversial, the Authors of this review aimed to assess CFX efficacy for patients with progressive MS. Among the pertinent literature, only five studies met the inclusion criteria of minimum methodological quality , with a total of 90 MS patients treated with CFX. 
 The data available were inadequate to attain all the objectives specified for the review. The Authors found poor evidence that CFX may slow progression in the medium‐term. They found also that side effects such as alopecia, nausea, vomiting and menses interruption (amenorrhea) occurred at high frequency, while some evidences might suggest adverse effects also after two years .

Background

Multiple sclerosis (MS) is considered an autoimmune disease associated with immune activity against central nervous system antigens. Immunosuppression and immuno‐modulation are the mainstays of the therapeutic strategies for this disease. Cyclophosphamide (CFX) is an alkylant agent with cytotoxic and immunosuppressive effects (Calabresi 1991), used in the treatment of different malignancies as well as autoimmune diseases (Wegener's granulomatosis, Periarteritis nodosa, Lupus Erythematosus Systemic) (Boumpas 1992). It has also been reported to reduce the severity and prevent Experimental Allergic Encephalomyelitis, the animal model of MS (Paterson 1974). 
 The efficacy of CFX in patients with progressive or relapsing form of MS remains controversial (Schluep 1997, Weiner 2002). Different treatment schedules have been adopted: varying dosages, route of administration (i.e. oral, intravenous), duration (ranging from a few days to months) and association with other drugs (i.e. adrenocorticotropic hormone‐ACTH‐steroids or other immunosuppressive agents) or with plasma exchange, followed or not by pulse maintenance treatment. The side effects and potential long‐term toxicity have been reported (DeRidder 1998;Fraiser 1991; Sen 1991, Talar‐Williams 1996). 
 In order to evaluate the existing data on CFX efficacy for patients with progressive MS, we have undertaken a systematic review of all relevant randomised controlled trials (RCTs) on this subject.

Objectives

The primary objectives of this review were to assess: 
 (1) whether the CFX treatment in MS patients slows disease progression; 
 (2) whether the effect of CFX is modified by the addition of steroids or ACTH.

The secondary objectives were to assess: 
 (1) whether CFX treatment prevent clinical relapses in secondary progressive (SP) MS; 
 (2) whether treatment efficacy is related to clinical variables (such as primary versus SP course; disease duration; and duration of progressive phase); 
 (3) the safety and tolerability of the treatment.

Methods

Criteria for considering studies for this review

Types of studies

All RCTs evaluating the clinical effect of CFX treatment in MS patients.

Types of participants

Patients affected by clinical definite MS, according to specific criteria (McDonald 1977; Poser 1983) who are in the progressive phase (primary and secondary), according to Lublin's criteria (Lublin 1996).

Types of interventions

We included RCTs evaluating the following treatment schedules. 
 (1) CFX administered alone or in combination with steroids or ACTH (the use of other immunosuppressive or immunomodulating agents and plasma‐exchange was excluded). 
 (2) Accepted CFX treatment schedules. 
 (a) Intensive immunosuppression schedule or "induction phase" (high dose in few days). 
 (b) Prolonged immunosuppression (fractionated doses in months). 
 (c) Pulse therapy (boluses of high doses for one day serially administered over time). 
 An arbitrary cut‐off was considered for high versus low doses, since, to the best of our knowledge, no clear indication has been suggested. 
 (3) Comparison group using placebo, no treatment, or the same co‐intervention (ACTH or steroids) only.

Types of outcome measures

The main outcome criterion was clinical disability progression, defined as an increase of 0.5 point of disability (according to the Kurtzke disability scale ‐ Kurtzke 1983) for patients with baseline EDSS greater than or equal to 6.0 and 1 point of disability for basal EDSS less than or equal to 5.5, sustained for three months. However, in order to allow comparisons among trials other less stringent criteria were accepted, such as an increase of 0.5 to 1 point, irrespective of basal EDSS, confirmed or not. 
 From each treatment group, the following outcome measures have been extracted: 
 (1) number of worsened patients; 
 (2) number of patients stable or improved (decrease of 0.5 to 1 point from baseline EDSS value); 
 (3) differences of disability between treatment versus control groups, evaluated as differences in progression of disability , defined as differences between the mean change of disability (EDSS final less EDSS onset) between treatment and control groups; 
 (4) relapse rate and time to exacerbation in secondary progressive patients with superimposed relapses; 
 (5) number of patients with side effects (laboratory hematological abnormalities, urotoxicity, infectious diseases). Both major (requiring or prolonging the hospitalization) and minor side effects have been considered. 
 The first four outcome measures have been evaluated at different points in time, according to treatment duration of the studies and follow up, while the fifth outcome was reported as frequency during the follow‐up. 
 The outcomes have been considered separately for primary and secondary progressive MS, if reported.

Search methods for identification of studies

Electronic searches

We searched the Cochrane MS Group Trials Register (June 2006), The Cochrane Central Register of Controlled Trials (CENTRAL) "The Cochrane Library 2006, Issue 3")(Appendix 1), MEDLINE (January 1966 to June 2006)(Appendix 2), EMBASE(1988 to June 2006)(Appendix 3)

Searching other resources

We also contacted authors and well‐known experts in this field to identify additional or unpublished studies. 
 In addition, we handsearched journal, the reference lists of each full retrieved paper and conference abstracts on the main meetings on Multiple Sclerosis.

Data collection and analysis

Study selection 
 Two reviewers (LLM and NM) independently decided whether to include or exclude each identified study; independently extracted and summarized data using an ad hoc form and cross‐checked the extracted data. Any discrepancy has been resolved by discussion with a third reviewer (BWG). 
 
 Data extraction 
 We extracted the following data from each study: 
 (1) exclusion/inclusion criteria; 
 (2) number of patients analysed in each group; 
 (3) key baseline variables of each group at entry; 
 (4) primary and secondary outcome measures; 
 (5) description of treatment schedules (dose and timing of drug administration); 
 (6) duration of follow‐up in each group; 
 (7) number of patients excluded and lost to follow‐up in each group; and 
 (8) side effects (type and incidence in each treatment group). 
 We contacted authors for their original data to allow comparisons among studies according to the predefined outcome measures. 
 
 Quality assessment 
 We assessed trial quality by: 
 (1) method of randomization; 
 (2) blinding of patients, clinicians and outcome assessors; 
 (3) the reporting of withdrawals and drop outs; 
 (4) intention‐to‐treat analysis; 
 (5) evidence of a priori sample size calculation; 
 (6) adoption of specific criteria of diagnosis; 
 (7) adoption of specific criteria of outcome measures. 
 Two reviewers ( LLM, NM) independently assessed the methodological quality of the included studies, using the Jadad checklist (Jadad 1996). The scale's score ranges from zero (lowest possible score) to five (highest possible score), with two points for randomisation (defined/not defined = 2/1)and blinding (Double/single = 2/1) and one point if drop outs were less than 15%. Studies scoring less than three points were regarded as being of low methodological quality.

Statistical analysis 
 We entered data into Cochrane Review Manager software and analysed it using Cochrane MetaView. Results are presented with 95% confidence intervals. We reported summary statistics for dichotomous outcomes as relative risk, and for continuous outcomes as weighted mean differences. We pooled estimates using a fixed effect model and assessed statistical heterogeneity using the chi squared test. We also considered sensitivity analysis evaluating the quality of studies methodology (i.e. adoption of blindness, specific criteria of diagnosis and outcome) which might affect the overall results.

Results

Description of studies

Of the 461 identified references (from January 1966 to June 2006), we selected 70 potentially relevant studies. Of those, 54 were rejected, primarily because of study design limitations (i.e. not controlled or randomised or preliminary studies). Of the 16 remaining RCTs, another eleven were excluded because the type of participants or the treatment schedules did not correspond to the inclusion criteria (D'Andrea 1991; Khatri 1984; Killian 1988;Weiner 1989; Weiner 1993, Smith 2004 as preliminary results and its final report Smith 2005) , one was another publication relevant to Weiner 1989 (Weiner 1990) and one was a comparative randomised trial including a third group, which was not treated and not randomised (Goodkin 1987). The protocol of another RCT has been reported ( Patti 2004), but the final study has been not published yet (see "Table of excluded studies"). Five studies were included: two, evaluating CFX efficacy versus placebo (Likosky 1991; The CCMSSG 1991), were used for the main analysis; three "comparative trials" (Hauser 1983; Wender 1988 and its extension Wender 1989) were analysed separately to evaluate the efficacy of ACTH co‐intervention (see "Table of included studies"). All of the included trials have randomised progressive patients with or without superimposed relapses (without clear distinction between secondary or primary progressive MS) (see additional Table 1 and Table 2). All evaluated the efficacy of CFX administered as an "intensive/induction" regimen, either alone (Likosky 1991) or associated or ACTH (Hauser 1983;Wender 1988) or prednisone (The CCMSSG 1991). The CCMSSG 1991 included a third arm, evaluating a different regimen as add on PE, this arm was excluded from the main analysis, in order to achieve better homogeneity. Finally, because the studies accepted for analysis adopted different definitions of treatment failure (see additional Table 2), the individual EDSS values at each time point were necessary for analysis. The EDSS data were reported in the full paper of Likosky 1991. We contacted Noseworthy for original data, (The CCMSSG 1991) which were promptly received and submitted for inclusion in the proposed meta‐analysis. According to the predefined outcome criteria, we were able to analyse the results at three time points: 12, 18 and 24 months. The studies comparing the efficacy of CFX plus ACTH versus ACTH alone were considered in a separate sub‐analysis: the available data allowed only one comparison (number of patients worsened of at least one EDSS point, at 12 months).

1. CFX versus placebo or no treatment.

reference	Likosky	Likosky	CCMSSG	CCMSSG
	CFX	Control	CFX	Control
No. patients	22	20	55	54
Age	43.9	40.8	nr	nr
Duration disease	9.2	11.5	9,2	10.4
Type MS
CP	+/‐ relapse	+/‐relapse	60%	50%
SP	unclear	unclear	40%	50%
EDSS entry (SD or range)	5.8(3‐8)	5.8(2‐7)	5.8+/‐0.6	5.8+/‐0.6
Prog rate ( EDSS/disease duration)	0.6	0.5	0.6	0.5

2. Comparative studies.

Reference	Wender	Wender	Hauser	Hauser
	CFX+ACTH	ACTH	CFX+ACTH	ACTH
No. patients	16	16	20	20
Age	40.2	41.6	32.9	35.2
Disease duration	10.4	10.7	(7.5)	(6)
Type of MS
CP	28%	28%	+ /‐relapse	+/‐relapse
SP	72%	72%	‐	‐
EDSS	4.8	4.9	5.8	5.6
Prog rate	0.5	0.4	0.8	0.9

Risk of bias in included studies

The assessment of trials quality is reported in the additional Table 3. Four included studies differ in trial design: three ( Hauser 1983, two centres;Likosky 1991‐ 13 centres; The CCMSSG 1991‐ nine centres were multicenter. One study was described as (The CCMSSG 1991) double‐blind, two (Likosky 1991;Wender 1988) as single‐blind and one (Hauser 1983) as an open study . The method of blinding was described in only one study (The CCMSSG 1991) in which, despite its description as double‐blind, most of the patients correctly identified the assigned therapy (Noseworthy 1994). The method of randomisation was specified in another study (The CCMSSG 1991). All the included studies adopted definite criteria of MS diagnosis and outcome measures (treatment failure). The definition of the type of MS in terms of Chronic Progressive (CP) and Secondary Progressive (SP) was unclear; it was reported as baseline data for two studies(The CCMSSG 1991; Wender 1988) but not considered further in the analysis. The quality assessment score for The CCMSSG 1991 was four, for Wender 1988 three, and or Hauser 1983 and Likosky 1991 two.

3. Trials quality.

Reference	score	Randomiz.	Blindness	Drop out
CCMSSG	4	Defined	double	9/168
		(2)	defined	55% completed 24 months
			(2)	(0)
Likosky	2	Not defined	single	7/42 (17%)
		(1)	(1)	(0)
Hauser	2	Not defined	Open	0/40
		(1)	(0)	(1)
Wender	3	Not defined	double	4/32 (10.5%)
		(1)	Not defined	(1)
			(1)

Effects of interventions

We were able to partially achieve the objectives outlined for this review . The available data enabled us to verify the efficacy of only one treatment schedule (intensive immunosuppression). According to the main outcome criteria (clinical disability progression) no effect was detected. Intensive immunosuppression with CFX (alone or associated with prednisone) in patients with progressive MS compared to placebo (152 participants) does not prevent the long‐term risk to clinical disability progression, as defined by the evolution to a next step of disability score (that is worsening of 0.5/1 EDSS points, according to baseline EDSS): at 12 months relative risk (RR) 0.92 (95% confidence interval (CI) 0.61 to 1.40); at 18 months RR 0.87 (95% CI 0.62 to 1.24); at 24 months RR 1.03 (95% CI 0.77 to 1.39) (see "Table of comparisons "01.CFX versus placebo. No. worsened"). We found no differences in the number of improved or stable patients; at 18 months only, the RR approximates the significance value (p= 0.09). However, the differences of the mean change of disability (final disability subtracted from the baseline) significantly favored the treatment group at 12 months (weighted mean difference (WMD) ‐ 0.21; 95% CI ‐ 0.25 to ‐ 0.17) and 18 months (WMD ‐ 0.19; 95% CI ‐ 0.24 to ‐ 0.14). At 24 months the increase of progression of disability was significantly worse in the CFX treated group (WMD +0.14; 95% CI 0.07 to 0.21) (see "Table of comparisons "01.CFX versus placebo; EDSS change"). It was not possible to determine the effect on the relapse rate in secondary progressive MS, due to the lack of data.

We performed a sub‐analysis to evaluate the benefit of ACTH co‐intervention, but the results must be interpreted with caution due to the heterogeneity of the studies (p = 0.0082) and the type of comparison (number of worsened patients at 12 months), (RR 0.36; 95% CI 0.19 to 0.67) (see "Table of comparisons "02. CFX+ACTH versus ACTH").

The side effects are showed in additional Table 4: only descriptive analysis was possible. Among the 90 CFX‐treated participants, the following main side effects were reported: alopecia, occurring in 100%, nausea and vomiting in 55 to 71%; amenorrhea in 42% (24% permanent) and cystitis in 4% of the cases. Major infections, like sepsis, or bronchopneumonia were reported in 11% of the participants. It should be mentioned that CFX was administered with particular procedures to reduce urotoxicity (hydration and sometimes catheterization) and nausea and vomiting (anti‐emetic drugs). CFX therapy also was reported to induce leukopenia: decrease of white blood cells (lowest 774), occurred five to seven days after the last dose, and began to recover seven days later. Five patients died: one related to MS at 18, one with lung cancer at 20, one with myocardial infarction at 15, one with bronchopneumonia at six months from onset of therapy, and one placebo treated for liver disease.

4. CFX side effects.

Author	CCMSSG	CCMSSG	Likosky	Likosky	Relative Risk
Group	CFX	Control	CFX	Control
Death	1 (pneumonia)	1 (hepatic failure)	1 (tumor)
			1 (MS)
			1 (cardiac failure)
Amenorrhea	42%	11%	nr	nr	3%
	24% permanent	7 % permanent
Alopecia	100%	16%	100%	nr	5%
			transient
Nausea	nr	nr	73%	nr
Cystitis	4%	nr	nr	nr
Sepsy	11%	nr	nr	nr

Discussion

Our systematic review, planned to evaluate the benefit of CFX therapy in patients with progressive MS, was only able to answer to a limited extent the a priori defined aims. The difficulties that we encountered were mostly related to the small number of CFX RCTs available (only four met the inclusion criteria), and to the heterogeneity of treatment regimens and study conducts. Although all the studies used the universally accepted Kurtzke EDSS disability scale (Kurtzke 1983), the outcome criteria (treatment failure) were differently defined . We were able to obtain raw data from the original studies, and so we were able to overcome these limitations by adopting a common outcome criterion of disease progression (defined as an increase of 0.5 point of disability for patients with baseline EDSS ≥ 6 and 1 point for EDSS ≤ 5.5 and the mean change of disability), pertinent for the kind of MS patients included in the studies. In this way, we were able to satisfy only the first main objective (whether the CFX treatment in MS patients slows disease progression) for only one treatment schedule (Induction).

The results indicated that CFX treatment did not prevent the evolution of disease to a higher EDSS status. The increase of disability (EDSS mean change) was significantly lower in the medium term in treated patients. This effect may be worthwhile for a short‐term therapy (such as a few days) in disabled MS patients However, this outcome is less sensitive and is based not on continuous score, which changes in multiples of 0.5. A significant increase of disability was found at 24 months, probably due further deterioration of the disease .

The lack of data unable us to verify the efficacy of other treatment schedules (such as boluses) and whether treatment efficacy is related to clinical variables, such as the form of disease (secondary versus primary progressive) or disease duration. The safety of the analysed schedule (intensive immunosuppression) was poor.

Authors' conclusions

Implications for practice.

This meta‐analysis did not allow definite conclusions on the efficacy of CFX therapy in MS patients, due to limited available data, poor quality of the studies, and the use of obsolete outcome criteria. The available data did not allow us to verify the efficacy of different schedules, in different form of the disease and in patients with different clinical characteristics (disability, disease duration, progression duration). The overall results show that CFX administered on an intensive schedule (the only analysed schedule) in disabled progressive patients has a clinical effect and safety profile that do not justify its use in clinical practice.

Implications for research.

CFX has been used in MS therapy since 1966 (Aimard 1966), and continue to be administered (Gauthier 2003,Gobbini 1999,Khan 2001, Perini 2003,Smith 2003, Zefir 2003), confirming the results of the international questionnaire on the clinical practice of immunosuppression treatments in MS (Hommes 2004 ). The low number of well‐conducted studies underlines the difficulty of properly evaluating the efficacy of CFX therapy (Noseworthy 1994;Weiner 2002) in patients with progressive form of the disease. This review shows that detailed and clear information on clinical characteristics of patients, on MS form (either secondary or primary progressive) are necessary to evaluate the efficacy of MS therapies, as well as blinding evaluation of treatment response (Noseworthy 1994'). A correct evaluation of treatment safety required a clear description of side‐effects. A systematic long ‐term surveillance of treated patients should be provided, because of the risk of potential toxicity, including cancer (Talar‐Williams 1996, DeRidder 1998, Fraiser 1991, Sen 1991). 
 The use of secondary outcome measures, including upper limb function and cognitive impairment (Rudick 2002), patient perception (Thompson 2002) and paraclinical surrogate measures (MRI), as well as blinding evaluation of treatment response ( Noseworthy 1994) are recommended for the adequate assessment of therapeutic effect. 
 It should be underlined that the included studies have been performed more than ten years ago. However, the use of CFX in the recent years presents two main differences in respect to the past, in both pharmacological and clinical aspects. High dose CFX (intensive induction schedule) is limited to patients undergoing stem cell transplantation ( Mc Guire 2003) , while at present the most commonly used regimen is every 4 to 8‐week pulse therapy given with or without corticosteroids, whose safety and tolerability have been recently reported (Portaccio 2003). Most of the treated patients are affected by rapidly deteriorating SP MS ( Perini 2003; Zefir 2003; Perini 2006 ) or RR refractory MS (Smith 2005; Reggio 2005; Gladstone 2006), excluding those in the later stages of the disease, when probably irreversible degenerative damage predominates over inflammation (Weiner 2002). Predictive factors of response have been identified, such as short progression duration (Delmont 2004; Zephir 2004); furthermore MRI data confirming the clinical response are reported (Smith 2005; Patti 2004 a; Gladstone 2006). However, none of these last studies fits the inclusion criteria of our analysis, preventing a critical analysis of these points.

What's new

Date	Event	Description
22 August 2008	Amended	Converted to new review format.

History

Protocol first published: Issue 4, 2000
 Review first published: Issue 4, 2002

Date	Event	Description
15 November 2006	New citation required and conclusions have changed	Substantive amendment
29 June 2006	New search has been performed	Searches were re run

Appendices

Appendix 1. CENTRAL search strategy

#1        "multiple sclerosis"

#2        MeSH descriptor Multiple Sclerosis explode all trees

#3        "Demyelinating disease*"

#4        MeSH descriptor Demyelinating Diseases, this term only

#5        "transverse myelitis"

#6        MeSH descriptor Myelitis, Transverse, this term only

#7        "neuromyelitis optica"

#8        "optic neuritis"

#9        MeSH descriptor Optic Neuritis explode all trees

#10      "encephalomyelitis acute disseminated"

#11      MeSH descriptor Encephalomyelitis, Acute Disseminated explode all trees

#12      "devic"

#13      MeSH descriptor Cyclophosphamide explode all trees

#14      cyclophosphamide

#15      phosphoramide AND mustards

#16      MeSH descriptor Phosphoramide Mustards explode all trees

#17      immunosuppression

#18      MeSH descriptor Immunosuppression explode all trees

#19      immunosuppressive AND agents

#20      autoimmune AND disease*

#21      MeSH descriptor Autoimmune Diseases explode all trees

#22      MeSH descriptor Antineoplastic Agents explode all trees

#23      antineoplastic

#24      antineoplastic AND agent*

#25      (#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11 OR #12)

#26      (#13 OR #14 OR #15 OR #16)

#27      (#17 OR #18 OR #19 OR #20 OR #21 OR #22 OR #23 OR #24)

#28      (#25 AND ( #26 AND #27 ))

Appendix 2. MEDLINE (PubMed) search strategy

((("Multiple Sclerosis"[mh]) OR ("Myelitis, Transverse"[mh:noexp]) OR ("Demyelinating Diseases"[mh:noexp]) OR ("Encephalomyelitis, Acute Disseminated"[mh:noexp]) OR ("Optic Neuritis"[mh])) OR ((("multiple sclerosis") OR ("neuromyelitis optica") OR ("transverse myelitis") OR (encephalomyelitis) OR (devic) OR ("optic neuritis")) OR ("demyelinating disease*") OR ("acute disseminated encephalomyelitis"))) AND (((randomized controlled trial[pt]) OR (controlled clinical trial[pt]) OR (randomized[tiab]) OR (placebo[tiab]) OR (drug therapy[sh]) OR (randomly[tiab]) OR (trial[tiab]) OR (groups[tiab])) NOT ((animals[mh]) NOT ((animals[mh]) AND (human[mh])))) AND ((("Cyclophosphamide"[Mesh]) OR ("Phosphoramide Mustards"[Mesh]) OR (cyclophosphamide) OR (phosphoramide AND mustards)) AND (((immunosuppression) OR ("Immunosuppression"[Mesh])) OR ((autoimmune AND disease) OR ("Autoimmune Diseases"[Mesh])) OR (("Antineoplastic Agents"[Mesh]) OR (antineoplastic) OR (antineoplastic AND agent*)) OR (immunosuppressive AND agents)))

Appendix 3. EMBASE (EMBASE.com) search strategy

((('encephalomyelitis'/exp) OR ('demyelinating disease'/exp) OR ('multiple sclerosis'/exp) OR ('myelooptic neuropathy'/exp) OR ('multiple sclerosis':ti,ab) OR ('neuromyelitis optica':ab,ti) OR (encephalomyelitis:ab,ti) OR (devic:ti,ab)) AND (('crossover procedure'/exp) OR ('double blind procedure'/exp) OR ('single blind procedure'/exp) OR ('randomized controlled trial'/exp) OR (random*:ab,ti) OR (factorial*:ab,ti) OR (crossover:ab,ti) OR (cross:ab,ti AND over:ab,ti) OR (placebo:ab,ti) OR ('double blind':ab,ti) OR ('single blind':ab,ti) OR (assign*:ab,ti) OR (allocat*:ab,ti) OR (volunteer*:ab,ti)) AND ((('cyclophosphamide'/exp) OR ('phosphoramide mustard'/exp) OR (cyclophosphamide:ab,ti) OR (phosphoramide:ab,ti AND mustards:ab,ti)) AND (((immunosuppression:ab,ti) OR ('immunosuppressive treatment'/exp)) OR (immunosuppressive:ab,ti AND agents:ab,ti) OR ((autoimmune:ab,ti AND disease*:ab,ti) OR ('autoimmune disease'/exp)) OR (('antineoplastic agent'/exp AND [embase]/lim) OR (antineoplastic:ab,ti AND agent*:ab,ti) OR (antineoplastic:ab,ti))))) AND [humans]/lim AND [embase]/lim

Data and analyses

Comparison 1. CFX versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 N° worsened at 12 months	2	152	Risk Ratio (M‐H, Fixed, 95% CI)	0.92 [0.61, 1.40]
2 N° Worsened at 18 months	2	133	Risk Ratio (M‐H, Fixed, 95% CI)	0.87 [0.62, 1.24]
3 N° Worsened at 24 months	2	112	Risk Ratio (M‐H, Fixed, 95% CI)	1.05 [0.79, 1.41]
4 N° ameliorated or stable at 12 months	2	152	Risk Ratio (M‐H, Fixed, 95% CI)	1.03 [0.77, 1.37]
5 N° ameliorated or stable at 18 months	2	133	Risk Ratio (M‐H, Fixed, 95% CI)	1.46 [0.95, 2.25]
6 N° ameliorated or stable at 24 months	2	112	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.62, 1.90]
7 Mean EDSS change at 12 months	2	138	Mean Difference (IV, Fixed, 95% CI)	‐0.21 [‐0.25, ‐0.17]
8 Mean EDSS change at 18 months	2	112	Mean Difference (IV, Fixed, 95% CI)	‐0.19 [‐0.24, ‐0.14]
9 Mean EDSS change at 24 months	2	96	Mean Difference (IV, Fixed, 95% CI)	0.14 [0.07, 0.21]

1.1. Analysis.

Comparison 1 CFX versus placebo, Outcome 1 N° worsened at 12 months.

1.2. Analysis.

Comparison 1 CFX versus placebo, Outcome 2 N° Worsened at 18 months.

1.3. Analysis.

Comparison 1 CFX versus placebo, Outcome 3 N° Worsened at 24 months.

1.4. Analysis.

Comparison 1 CFX versus placebo, Outcome 4 N° ameliorated or stable at 12 months.

1.5. Analysis.

Comparison 1 CFX versus placebo, Outcome 5 N° ameliorated or stable at 18 months.

1.6. Analysis.

Comparison 1 CFX versus placebo, Outcome 6 N° ameliorated or stable at 24 months.

1.7. Analysis.

Comparison 1 CFX versus placebo, Outcome 7 Mean EDSS change at 12 months.

1.8. Analysis.

Comparison 1 CFX versus placebo, Outcome 8 Mean EDSS change at 18 months.

1.9. Analysis.

Comparison 1 CFX versus placebo, Outcome 9 Mean EDSS change at 24 months.

Comparison 2. CFX and ACTH versus ACTH.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 No. worsened ( > 1) EDSS at 12 months	2	72	Risk Ratio (M‐H, Fixed, 95% CI)	0.36 [0.19, 0.67]

2.1. Analysis.

Comparison 2 CFX and ACTH versus ACTH, Outcome 1 No. worsened ( > 1) EDSS at 12 months.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Hauser 1983.

Methods	controlled randomized open
Participants	clinically definite progressive multiple sclerosis
Interventions	(1) CFX IV 400‐500 mg /day (10‐14) (5‐7 gr)+ACTH versus (2) CFX os 2 mg/Kg/day (42) (5 gr) + ACTH+PE versus (3) ACTH IV 25 U tapering for 21 days WBC flag
Outcomes	EDSS Ambulation Index
Notes	Time points: 6,12 months
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Likosky 1991.

Methods	controlled randomized single blind
Participants	definite progressive multiple sclerosis
Interventions	(1)CFX IV 400‐500 mg/day, 5 x/week ( tot 4.6 ) versus (2) folic acid WBC flag
Outcomes	EDSS Ambulation Index Incapacity Status Environmental Status
Notes	Timepoints: 12,18,24 months
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

The CCMSSG 1991.

Methods	controlled randomized double blind
Participants	definite progressive multiple sclerosis
Interventions	(1)CFX IV 1 gr AD +prednisone 40 mg tapered for 16 days) (tot < =9gr) versus (2) CFX os(1.5‐2 mg/kg/day for 22 weeks (18gr) +prednisone+PE versus (3) placebo WBC flag
Outcomes	EDSS
Notes	Time points: 6,12,18,24,30,36 months
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Wender 1988.

Methods	Controlled randomized single blind
Participants	chronic progressive multiple sclerosis
Interventions	(1) CFX IV 200 mg /day for 20 days (tot 4 gr) + ACTH im versus (2) ACTH
Outcomes	DSS
Notes	Time points: 6,12,18,24 months
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

(*) control group was not randomized 
 ACTH: adrenocorticotropic hormone 
 CFX: Cyclophosphamide 
 DSS: Disability Status score 
 EDSS: Expanded Disability Status Scale 
 im: intra‐muscular 
 IV: intravenous 
 PE: Plasma Exchange 
 os: 
 WBC flag: White blood cells < 4000 therapy discontinuation (< 4,500 for CCMSSG)

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Aimard 1966	Design case report
Carter 1988	Design (not controlled)
Cendrowski 1973	Design (comparative schedule, not randomized, Type of schedule)
D'Andrea 1990	Design (not controlled) Type of participants (relapsing‐remitting))
D'Andrea 1991	Type of participants (relapsing‐remitting)
Delmont 2004	Design ( not controlled)
Drachman 1975	Design (not controlled) Type of participants (patients in relapse)
Gauthier 2003	Preliminary report (Abstract) Design : Not controlled
Girard 1967	Design (not controlled)
Gladstone 2006	Design ( not controlled)
Gobbini 1999	Design (not controlled) Paraclinical end points (MRI)
Gonsette 1977	Design (not controlled) Type of participants (relapsing‐remitting)
Goodkin 1987	Comparative trials Control group was not randomized
Gopel 1972	Design and type of treatment (Including azathioprine)
Haerer 1991	Not full paper (abstract)
Hohol 1992	Not full paper (abstract)
Hohol 1999	Design (not controlled)
Hommes 1975	Design (not controlled)
Hommes 1980	Design (not controlled)
Huston 1983	Design (not controlled)
Khan 2001	Design (Not controlled)
Khatri 1984	Design (not controlled)
Khatri 1985	Design and treatment schedule (comparative trial evaluating the efficacy of plasma exchange in CFX treated patients)
Khatri 1991	Design (not controlled) and treatment schedule
Killian 1988	Type of participants (relapsing‐remitting MS patients)
Kornhuber 1986	Design (not controlled) Safety study
Kornhuber 1987	Design (not controlled) Safety study
La Mantia 1997	Design (not randomized) Treatment schedule (comparative trial)
Laterre 1989	Design (not controlled)
Lebrun 2006	Design (not controlled) Ooutcome criterai (fatigue)
Mancardi 2001	Type of treatment (autologous stem cell transplantation)
Manova 2000	Design (controlled not randomized)
Manova 2001	Design (controlled not randomized) Outcome measures
Mauch 1989	Design (not randomized)
Merico2002	Case report
Mickey 1987	Design (not controlled) Paraclinical end points (immunological)
Millac 1969	Design (unclear) Lack of clinical data
Millefiorini 1990	Design (not controlled)
Myers 1987	Design (not controlled) Paraclinical end points (immunological)
Paradowski 1999	Design (comparative study not randomized)
Patti 2001	Design (not controlled)
Patti 2004	Type of partecipants (relapsing‐remitting patients) Presentation of the protocol of a double‐blind, placebo‐controlled study No data
Patti 2004 a	Design of the study (open) Type of partecipants (relapsing remitting patients with "rapidly transitional form") Type of schedule (co‐intervention with interferon‐beta) Exrension of a previous study (Patti, J Neurol Neurosurg Psychiatry 2001)
Perini 2003	Design (not controlled)
Perini 2006	Design (controlled‐not randomized) Type of comparison (cyclophosphamide versus mitoxantrone)
Portaccio 2003	Design (not controlled)
Reggio 2005	Design (open non controlled study) Type of partecipants (relapsing‐remitting patients) Type of schedule (co‐intervention with interferon beta)
Rensel 1999	Design (not controlled)
Rivera 1988	Design (not controlled) Not full paper
Salmaggi 1994	Design (not controlled)
Siracusa 1987	Design (not controlled)
Smith 2003	Preliminary report (abstract) Type of schedule (including interferon beta)
Smith 2004	Type of partecipants ( relapsing‐remitting patients ) Type of schedule (including interferon beta 1 a) Randomized single‐blind controlled study Preliminary data
Smith 2005	Type of partecipants ( relapsing‐remitting patients ) Type of schedule (including interferon beta 1 a) Randomized single‐blind controlled study Final results of the study Smith 2004.
Trouillas 1989	Design (randomization unclear) Treatment schedule (Including azathioprine)
Weiner 1989	Type of participants (MS patients in relapse)
Weiner 1993	Type of participants (relapsing MS patients) and treatment schedules
Weinstock‐Guttman 97	Design (not controlled)
Zefir 2002	Design (not controlled)
Zefir 2003	Design (not controlled)
Zephir 2004	Design (not controlled)
Zephir 2005	Design (not controlled) Outcome mesaure (cognitive function)

Contributions of authors

Dr L La Mantia conceived the idea and development of the project, posed the questions of the research project and coordinated the review. Prof B Weinstock defined the final design of the protocol, the objectives of the study, the relevance and validity of the papers. They both analyzed the results and wrote the text of the review. Dr L La Mantia extracted the data, which were cross‐checked by Dr N Mascoli. B Incorvaia extracted the data in the previous version of the review . Dr L La Mantia analyzed the data. Prof B Weinstock reviewed the results and the final version of the study. Dr C Milanese wrote to Authors of trials for the required data and reviewed the manuscript drafts. Dr R D'Amico assisted in statistical analysis and helped with the preparation of the review.

Sources of support

Internal sources

• Fondazione I.R.C.C. S. Istituto Neurologico "Carlo Besta", Italy.

External sources

• No sources of support supplied

Declarations of interest

The review was assembled, analysed and reported independently of any company. Dr L La Mantia has received travel expenses from Dompè Biotec pharmaceutical company for participating at one neurological meeting.

Edited (no change to conclusions)