Scolaris Content Display Scolaris Content Display

اینترلوکین‐2 به عنوان درمان نگهدارنده برای کودکان و بزرگسالان مبتلا به لوکمی میلوئید حاد در اولین بهبودی کامل

Collapse all Expand all

چکیده

available in

پیشینه

لوکمی میلوئید حاد (acute myeloid leukaemia; AML) یک سرطان بدخیم در سلول‌های بنیادی خونساز است. درمان AML شامل دو مرحله درمانی است: فاز القای بهبودی برای رسیدن به بهبودی کامل و سریع (complete remission; CR) و فاز تثبیت برای رسیدن به بهبودی مولکولی بادوام. افرادی که CR را تجربه کرده‌اند، در معرض خطر عود AML هستند، چنین افرادی با عود AML، احتمال بقای ضعیفی دارند. بنابراین، نیاز مداوم به درمان‌هایی برای بهبود بیشتر پیش‌آگهی وجود دارد. اینترلوکین‐2 (Interleukin; IL‐2)، یک سیتوکین (cytokine) ایمنی محرک و یک جایگزین برای درمان استاندارد در افراد مبتلا به AML است که جهت حفظ اثربخشی آن پس از تثبیت درمان استفاده می‌شود. درمان نگهدارنده، بخش جدایی‌ناپذیر درمان استاندارد AML نیست. در این زمینه، هم‌چنین مطالعاتی به منظور ارزیابی اثربخشی IL‐2 به عنوان درمان نگهدارنده برای افراد مبتلا به AML که به اولین CR رسیده‌اند، انجام شده است، اما تاثیر IL‐2 هنوز به طور کامل شناسایی نشده است.

اهداف

این مطالعه به منظور بررسی اثربخشی و بی‌خطری IL‐2 به عنوان درمان نگهدارنده در کودکان و بزرگسالانی انجام شده که یا مبتلا به AML بوده و CR اولیه خود را به دست آورده بودند یا بیماری در آنها عود نکرده بود.

روش‌های جست‌وجو

در این مرور، پایگاه مرکزی ثبت کارآزمایی‌های بالینی کاکرین (CENTRAL) (کتابخانه کاکرین (Cochrane Library)؛ شماره 8، 2015)؛ MEDLINE (از سال 1950 تا آگوست 2015)؛ EMBASE (از سال 1950 تا آگوست 2015)؛ LILACS (از سال 1982 تا آگوست 2015)؛ CBM (از سال 1978 تا آگوست 2015)، خلاصه مقالات کنفرانس‌های مرتبط (از سال 2000 تا 2015)، و متارجیستری از کارآزمایی‌های کنترل شده برای کارآزمایی‌های در حال انجام و منتشر نشده (از ابتدا تا آگوست 2015)، را به طور سیستماتیک جست‌وجو کردیم. علاوه بر این، فهرست منابع کارآزمایی‌ها و مرورهای مرتبط را غربالگری کردیم.

معیارهای انتخاب

مطالعات واجد شرایط در این مرور، کارآزمایی‌های تصادفی‌سازی و کنترل شده‌ای (randomised controlled trials; RCTs) بود که برای مقایسه IL‐2 با عدم درمان، در افراد مبتلا به AML که برای اولین بار CR خود را به دست آورده بودند یا بیماری در آنها عود نکرده بود، صورت گرفته است. در این مرور، مطالعه‌ای که استفاده از IL‐2 را در برابر بهترین مراقبت حمایتی یا شیمی‌درمانی نگهدارنده مقایسه کرده باشد یا مطالعاتی که در آنها IL‐2 به همراه شیمی‌درمانی نگهدارنده در برابر شیمی‌درمانی به تنهایی با هم مقایسه شده باشند، یافت نشد.

گردآوری و تجزیه‌وتحلیل داده‌ها

دو نویسنده مرور به‌طور مستقل از هم، مطالعات را غربالگری کرده، داده‌ها را با استفاده از فرم استخراج از پیش تعریف شده استخراج و خطر سوگیری (bias) مطالعات وارد شده را بررسی کردند. داده‌های مربوط به پیامدهای زیر استخراج شد: بقای بدون بیماری، بقای کلی (overall survival; OS)، بقای بدون عوارض، مورتالیتی ناشی از درمان، عوارض جانبی و کیفیت زندگی. در این مطالعه، تاثیر درمان بر پیامدهای مربوط به زمان تا رخداد یک پیامد و پیامدهای دو‐حالتی، را به ترتیب با استفاده از نسبت خطر (HR) و خطر نسبی (RR) اندازه‌گیری کردیم. در این مطالعه هم‌چنین، به جز در مواردی که ناهمگونی معنی‌داری بین مطالعات وجود داشته باشد، از روش واریانس معکوس، برای ترکیب نسبت مخاطرات با استفاده از مدل اثر‐ثابت استفاده شد.

نتایج اصلی

این مرور شامل نه RCT با مجموع 1665 شرکت‌کننده است که در آنها بین IL‐2 و گروه عدم درمان مقایسه انجام شده بود. شش مطالعه شامل شرکت‏‌کنندگان بزرگسال و سه مطالعه نیز شامل شرکت‏‌کنندگان بزرگسال و کودک بود. با این حال، در سه مطالعه مربوط به بزرگسالان و کودکان، داده‌هایی درباره کودکان گزارش نشده بود، بنابراین در پژوهش حاضر امکان انجام تجزیه‌وتحلیل زیر‐گروه برای کودکان وجود نداشت. در یک مطالعه که در کشور چین انجام شده بود، هیچ یک از پیامدهای مورد نظر این مرور گزارش نشده بود. بنابراین در پژوهش حاضر، تعداد شش کارآزمایی شامل 1426 شرکت‌کننده در متاآنالیز (meta‐analysis) بقای بدون بیماری، و تعداد پنج کارآزمایی شامل 1355 شرکت‌کننده در متاآنالیز بقای کلی استفاده شد. شواهدی مبنی بر تفاوت معنی‌دار بین گروه IL‐2 و گروه عدم درمان، از نظر بقای بدون بیماری (HR: 0.95؛ 95% CI؛ 0.86 تا 1.06؛ 0.37 = P؛ شواهد با کیفیت پائین) یا بقای کلی (HR: 1.05؛ 95% CI؛ 0.95 تا 1.16؛ 0.35 = P؛ شواهد با کیفیت متوسط) یافت نشد. بر اساس اطلاعات یکی از کارآزمایی‌ها با تعداد161 شرکت‌کننده، IL‐2 هیچ تاثیری در بقای بدون عارضه نداشت (HR: 1.02؛ 95% CI؛ 0.79 تا 1.32؛ 0.88 = P؛ شواهد با کیفیت پائین). با توجه به اطلاعات حاصل از یک کارآزمایی با 308 شرکت‌کننده، عوارض جانبی (از جمله ترومبوسیتوپنی (thrombocytopenia)، نوتروپنی (neutropenia)، ضعف/خستگی، و عفونت/تب) در شرکت‏‌کنندگانی که IL‐2 دریافت کرده بودند، شایع‌تر بود. هیچ موردی از مورتالیتی به دلیل بروز عوارض جانبی گزارش نشده است. هیچ یک از مطالعات وارد شده، مورتالیتی ناشی از درمان یا کیفیت زندگی را گزارش نداده‌اند.

نتیجه‌گیری‌های نویسندگان

هیچ گونه شواهدی حاکی از تفاوت معنی‌دار بین گروه درمان نگهدارنده IL‐2 و گروه عدم درمان، از نظر بقای بدون بیماری یا بقای کلی، در افراد مبتلا به AML در اولین CR، یافت نشد، با این حال، کیفیت شواهد متوسط یا پائین است و به احتمال زیاد یا بسیار بالا، انجام پژوهش بیشتر، تاثیر مهمی بر تخمین یا اطمینان نسبت به این تخمین دارد. به نظر می‌رسد عوارض جانبی در شرکت‏‌کنندگانی که تحت درمان با IL‐2 قرار گرفته‌اند، شایع‌تر است، اما کیفیت شواهد بسیار پائین و سطح اطمینان در این تخمین بسیار نامطمئن است. بنابراین، پیش از آن که بتوان نتیجه‌گیری‌های قطعی در مورد این مسائل به دست آورد، انجام کارآزمایی‌های تصادفی‌سازی شده آینده‌نگر بیشتری مورد نیاز است.

PICOs

Population
Intervention
Comparison
Outcome

The PICO model is widely used and taught in evidence-based health care as a strategy for formulating questions and search strategies and for characterizing clinical studies or meta-analyses. PICO stands for four different potential components of a clinical question: Patient, Population or Problem; Intervention; Comparison; Outcome.

See more on using PICO in the Cochrane Handbook.

خلاصه به زبان ساده

نقش اینترلوکین‐2 به عنوان درمان نگهدارنده در کودکان و بزرگسالان مبتلا به لوکمی میلوئید حاد در اولین بهبودی کامل

پیشینه

لوکمی میلوئید حاد (acute myeloid leukaemia; AML)، یک سرطان بدخیم در سلول‌های خونی است. در اغلب افراد مبتلا به AML، برای از بین بردن سلول‌های سرطانی، شیمی‌درمانی یا پیوند مغز استخوان انجام می‌شود. در پاسخ به این درمان‌ها، ممکن است افراد به بهبودی کامل (complete remission; CR)، یعنی وضعیت بدون سلول‌های سرطانی در خون و علائم سرطان خون دست یابند. اگرچه این بیماری، به خوبی در آغاز درمان کنترل می‌شود، AML ممکن است بازگردد یا دوباره پیشرفت کند، در نتیجه در امید به زندگی، سبب نارضایتی شود. افراد پس از CR، هنوز هم نیاز به درمان بیشتر دارند. اینترلوکین (Interleukin)‐2 (یا به اختصار IL‐2) جایگزینی است که به عنوان درمان نگهدارنده برای حفظ تاثیر درمانی و جلوگیری از پیشرفت بیماری استفاده می‌شود. نتایج بالینی مربوط به مقایسه استفاده از IL‐2 و درمان نگهدارنده در افرادی که در آنها CR به دست آمده بود، منتشر شده، اما تاثیر IL‐2 همچنان نامشخص است.

سوال مطالعه مروری

هدف از این مرور، ارزیابی و خلاصه کردن مطالعات علمی است که بر اثربخشی و تاثیر مضر IL‐2 به عنوان درمان نگهدارنده در کودکان و بزرگسالان مبتلا به AML که برای بار اول به CR دست یافته‌اند، و بیماری در آنها عود نکرده، تمرکز کرده‌اند.

ویژگی‌های مطالعه

در این مرور، جست‌وجو در تمام بانک‌های اطلاعاتی برای یافتن مطالعات مرتبط که در فاصله بین ژانویه 1950 تا آگوست 2015 منتشر شده‌اند، انجام گرفت و 679 استناد شناسایی شد. تعداد نه مطالعه شامل 1665 شرکت‌کننده، وارد مطالعه شد. تعداد شرکت‌کنندگان در این کارآزمایی‌ها از 24 تا 528 نفر بین سال‌های 1991 و 2008 و با میانه (median) دوره پیگیری از 2.4 تا 8.3 سال متغیر است. سن شرکت‏‌کنندگان از 0 تا بیش از 60 سال بود. شرکت‏‌کنندگان در شش مطالعه، افراد بزرگسال بودند و سه مطالعه دیگر شامل هر دو مورد بزرگسالان و کودکان بود. با این حال، در این سه مطالعه، داده‌های مربوط به کودکان گزارش نشده است، بنابراین امکان انجام تجزیه‌وتحلیل زیر‐گروه برای کودکان وجود نداشت. در یک مطالعه که در کشور چین انجام شده بود، هیچ یک از پیامدهای مورد نظر این مرور گزارش نشده بود. منابع مالی این کارآزمایی‌ها، عمدتا توسط موسسات یا سازمان ملی تامین شده بود. شرکت‏‌کنندگان گروه درمان، درمان نگهدارنده IL‐2 را دریافت کرده و شرکت‏‌کنندگان گروه کنترل، هیچ درمان دیگری دریافت نکردند. هیچ مطالعه‌ای که در آن IL‐2 را با شیمی‌درمانی نگهدارنده مقایسه کرده باشد یا IL‐2 به همراه شیمی‌درمانی نگهدارنده را با شیمی‌درمانی نگهدارنده به تنهایی مقایسه کرده باشد، شناسایی نکردیم. شواهد تا آگوست 2015 به‌روز است.

نتایج کلیدی

در این مرور داده‌ها را با توجه به تاثیر IL‐2 بر بقای بدون بیماری (فاصله زمانی از تاریخ بهبودی تا عود لوکمیک یا مرگ‌ومیر به هر علتی) و بقای کلی (مدت زمان بین ورود شرکت‏‌کنندگان به مطالعه و زمان فوت آنها) در مقایسه با عدم درمان، آنالیز کردیم. هیچ گونه شواهدی برای تفاوت در بقای بدون بیماری یا بقای کلی بین گروه IL‐2 و گروه کنترل نیافتیم. پیامد دیگر در نظر گرفته شده، بقای عاری از رویداد (فاصله زمانی بین تاریخ تصادفی‌سازی یا ورود به مطالعه تا شکست درمان، عود اول، یا مرگ‌ومیر به هر علتی) بود، یافته‌ها نشان داد که IL‐2 بقای بدون عارضه را افزایش نمی‌دهد. هیچ یک از مطالعات وارد شده، شامل آنالیز مربوط به مورتالیتی ناشی از درمان یا کیفیت زندگی نبود. علاوه بر این وجود عوارض جانبی مربوط به درمان IL‐2، بررسی شد. نتایج به دست آمده بر اساس شواهد با کیفیت بسیار پائین نشان داد که عوارض جانبی در شرکت‏‌کنندگان دریافت کننده IL‐2 نسبت به عدم درمان، شایع‌تر است. عوارض جانبی گزارش شده شامل کاهش تعداد پلاکت خون یا تعداد نوتروسیت (neutrocyte)، ضعف، خستگی، تب و عفونت بود. هیچ موردی از مرگ‌ومیر مربوط به عوارض جانبی، گزارش نشده بود.

کیفیت شواهد

کیفیت شواهد با توجه به بقای بدون بیماری و بقای بدون عارضه، پائین بود. کیفیت شواهد برای بقای کلی در سطح متوسط و برای عوارض جانبی در سطح بسیار پائین بود.

نتیجه‌گیری

هیچ گونه شواهدی مبنی بر تفاوت بین گروه نگهدارنده درمان IL‐2 و گروه عدم درمان، از نظر بقای بدون بیماری یا بقای کلی در افراد مبتلا به AML در اولین CR یافت نشد، با این حال، کیفیت شواهد در سطح متوسط یا پائین است و انجام پژوهش بیشتر به احتمال بسیار زیاد، تاثیر مهمی در تخمین یا سطح اطمینان به تخمین به دست آمده خواهد داشت. به نظر می‌رسد عوارض جانبی در افراد تحت درمان با IL‐2 شایع‌تر است، اما کیفیت شواهد در سطح بسیار پائین و سطح اطمینان نسبت به تخمین‌ها بسیار نامطمئن است. بنابراین، پیش از آن که بتوان نتیجه‌گیری‌های قطعی در مورد این مسائل به دست آورد، انجام کارآزمایی‌های تصادفی‌سازی شده آینده‌نگر بیشتری مورد نیاز است.

Authors' conclusions

Implications for practice

We found no evidence for a difference between IL‐2 maintenance therapy and no treatment with respect to disease‐free survival or overall survival of people with AML in first CR. Adverse events seemed to be more frequent in participants treated with IL‐2. Thus, the results of the review do not support the use of IL‐2 as maintenance therapy in people with AML in first CR.

Implications for research

  1. As mentioned in the Results section, the existing results of outcomes of interest were mainly based on Western studies. Ethnicity has been described as an independent prognostic factor in long‐time survival of people with AML, and the survival prospects varied in people with AML in different ethnic groups (Alcalai 2003; Aplenc 2006). Whether the efficacy of IL‐2 varies with ethnicity remains unknown and warrants further research.

  2. Limited studies reported data on event‐free‐survival, treatment‐related mortality, adverse events, and quality of life, so the efficacy of IL‐2 on these outcomes should be further investigated.

  3. As the quality of the evidence for the primary outcomes is low, further research is required, which could affect the estimates and confidence.

Summary of findings

Open in table viewer
Summary of findings for the main comparison.

IL‐2 compared with no treatment for people with AML in first complete remission

Patient or population: People with AML in first complete remission

Settings: Maintenance therapy

Intervention: IL‐2

Comparison: No treatment

Outcomes

Illustrative comparative risks[1]

Relative effect
(95% CI)

No of Participants
(studies)

Quality of the evidence
(GRADE)

Comments

Assumed risk

Corresponding risk

No treatment

IL‐2

Relapses/death (instead of disease‐free survival)[2]

Follow‐up at median 5 years

590 per 1000

571 per 1000

(535 to 611)

HR 0.95 (0.86 to 1.06)

1426
(6 studies)

⊕⊕⊝⊝[3]
Low

Mortality (instead of overall survival)[2]

Follow‐up at median 5 years

480 per 1000

497 per 1000

(463 to 532)

HR 1.05 (0.95 to 1.16)

1355
(5 studies)

⊕⊕⊕⊝[4]
Moderate

Death/progress (instead of event‐free survival)[2]

Follow‐up at 5 years

280 per 1000

285 per 1000

(229 to 352)

HR 1.02 (0.79 to 1.32)

161

(1 study)

⊕⊕⊝⊝[3]
Low

Treatment‐related mortality

See comments

See comments

Not estimable

0

(0)

See comments

None of the studies reported treatment‐related mortality

Adverse events: hypersensitivity

Follow‐up at median 6 years

See comments

See comments

RR 17.13 (0.99 to 295.26)

528

(1 study)

⊕⊝⊝⊝[3,5]
Very low

Illustrative comparative risks are not estimable as the assumed risk is zero

Adverse events: fatigue

Follow‐up at median 6 years

10 per 1000

71 per 1000

(21 to 233)

RR 7.05 (2.13 to 23.36)

528

(1 study)

⊕⊝⊝⊝[3,5]
Very low

Adverse events: rigor/chills

Follow‐up at median 6 years

See comments

See comments

RR 33.25 (2.01 to 551.36)

528

(1 study)

⊕⊝⊝⊝[3,5]
Very low

Illustrative comparative risks are not estimable as the assumed risk is zero

Adverse events: arthralgia/myalgia

Follow‐up at median 6 years

See comments

See comments

RR 19.14 (1.12 to 327.24)

528

(1 study)

⊕⊝⊝⊝[3,5]
Very low

Illustrative comparative risks are not estimable as the assumed risk is zero

Quality of life

See comments

See comments

Not estimable

0

(0)

See comments

None of the studies reported quality of life

CI: confidence interval; HR: hazard ratio; IL‐2: interleukin‐2; RR: risk ratio

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

1. The basis for the assumed risk is the median control group risk across studies (The absolute effects come directly from the included study if only one study was included). The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). The HRs are first converted to RRs and the corresponding risks are then calculated from the RRs. Corresponding intervention risk, per 1000 people = 1000 × assumed control risk× RR.

2. Absolute effects were estimated from the HRs, and both are reported in the same row.

3. Downgrading two points for study limitations (high risk of detection bias due to lack of blinding; sequence generation and allocation concealment in some studies were not reported, leading to potential selection bias).

4. Downgrading one point for study limitations (sequence generation and allocation concealment in some studies were not reported, leading to potential selection bias).

5. Downgrading one point for imprecision (wide confidence interval).

Background

Description of the condition

Acute myeloid leukaemia (AML) is a malignant cancer of hematopoietic stem cells characterised by the rapid growth of non‐functional myeloblasts in the bone marrow and interference with the production of normal blood cells. The symptoms of AML are related to the replacement of normal bone marrow with leukaemic cells, which will lead to functional pancytopenia, with the symptoms of anaemia, neutropenia, and thrombocytopenia (Lowenberg 1999). AML is a relatively uncommon cancer, with an estimated incidence of 3.5 per 100,000, which has remained stable during recent decades (Howlader 2013). The incidence of AML increases with age, and people over 60 years make up a large proportion of those with this disease, and their survival time is only half of that of younger patients (Buchner 2009).

There are two commonly used classifications for AML: the French‐American‐British (FAB) system and the World Health Organization (WHO) system. The former classification system was introduced in 1976 and categorises AML into eight subtypes, from M0 through to M7, using morphological and cytochemical methods. It sets the threshold between high‐grade myelodysplastic syndromes and AML greater than or equal to 30% blasts (Bennett 1976). The WHO classification system was introduced in 1999 and revised in 2008. It is based on morphology and newer prognostic factors to categorise AML into 17 subtypes and lowers the threshold to greater than 20% blasts in diagnosis (Harris 1999; Vardiman 2009).

Normally, the treatment of AML consists of two well‐validated treatment phases: the remission induction phase and the consolidation (postremission) phase. The remission induction phase aims to achieve a rapid, complete remission (CR), and the consolidation phase aims to achieve a durable molecular remission. Possibilities for consolidation are autologous stem cell transplantation (auto‐SCT) and allogenic stem cell transplantation (allo‐SCT), as well as chemotherapy (Heitger 2002). The effect of IL‐2 in people with AML is related to many prognostic factors, such as the person's age, performance status, karyotype and cytological AML subtype (Appelbaum 2006; Frohling 2006; Juliusson 2009; Wheatley 2009). Ethnicity is also an independent predictor for long‐term survival in adults and children with AML (Alcalai 2003; Aplenc 2006).

Description of the intervention

People with AML who achieve CR and subsequently relapse have a poor survival prospect (Rowe 2005). After the consolidation period in AML treatment, maintenance therapy is required, the aim of which is to maintain patients in first CR and prevent relapse (Krug 2010). Maintenance therapy is not an integral part of the standard treatment for AML. The effect of interleukin‐2 (IL‐2) as remission maintenance treatment in people with AML remains controversial. Case reports and previous small clinical trials have shown that IL‐2 induced prolonged remissions and may bring a modest benefit to people with AML for remission maintenance in first CR (Farag 2002; Maraninchi 1991; Meloni 1994; Stein 2002). A series of randomised controlled trials (RCTs) have demonstrated the putative benefits of IL‐2 in survival outcomes in people with AML (Baer 2008; Blaise 2000; Kolitz 2014; Lange 2008; Pautas 2010). However, the trial by Kolitz (Kolitz 2014), which evaluated the efficacy of IL‐2 in 214 people with AML in first CR after completing all planned chemotherapy, showed no significant difference between the IL‐2 treatment group and the no‐treatment group in terms of the three‐year disease‐free survival rate (56% versus 45%; P = 0.11) and three‐year overall survival rate (68% versus 61%; P = 0.09). A recent individual patient data meta‐analysis suggested that IL‐2 alone as a remission maintenance therapy was not effective for people with AML in first CR (Buyse 2011). Only the result of a randomised phase 3 trial with 320 people with AML in CR indicates that IL‐2 in combination with histamine dihydrochloride can significantly reduce the relapse risk (Brune 2006).

In addition to serving as maintenance therapy, the potential role of low‐dose IL‐2 as part of consolidation immunotherapy for the maintenance of CR in AML is obvious, particularly to people in CR with a low or minimal burden of leukaemia (Thoren 2009).

How the intervention might work

IL‐2 is a strongly immune‐stimulatory cytokine, which significantly promotes the proliferation of T lymphocytes responsible for tumour‐specific cytotoxic function, and augments the cytotoxic effect of natural killer cells (T and NK cells) (Lauria 1994), suggesting that IL‐2 is a potential candidate for leukaemia therapy by its cytotoxic lymphocytes pathway. Previous preclinical studies have proved that IL‐2 alone was effective in treating leukaemia in animal models (Fierro 1988; Johnson 1989). A few non‐randomised trials also reported that high doses of IL‐2 contributed to achieving objective remissions in patients (Maraninchi 1991; Maraninchi 1998). IL‐2 can also replicate the beneficial effects of allogeneic hematopoietic stem cell transplantation to prevent leukaemic relapse. However, IL‐2 treatment at high doses had considerable adverse effects such as fever, flushing, vomiting, diarrhoea, fatigue, thrombocytopenia, renal dysfunction, and pulmonary oedema (Lauria 1994; Stoppa 1991; Tajima 1996).

Consequently, low‐dose IL‐2 therapy was considered for leukaemic maintenance therapy. Treatment with low‐dose IL‐2, which saturates only high‐affinity IL‐2 receptors, resulted in a selective expansion of a subset of NK cells, especially CD56bright and CD16di(Caligiuri 1993). In addition, the effect of antigen‐independent cytotoxicity against NK cell‐resistant AML blasts achieved maximal activation level when these high‐affinity receptors were saturated by low‐dose IL‐2. A study also demonstrated that low‐dose IL‐2 as postremission therapy was well tolerated when administrated to elderly people with AML (Farag 2002).

Why it is important to do this review

To date, several RCTs have evaluated IL‐2 as remission maintenance therapy in people with AML (Baer 2008; Blaise 2000; Kolitz 2014; Lange 2008; Willemze 2009). Given the difference in age, length of follow‐up, remission induction regimens, dose of IL‐2, and sample size, no consensus has been reached on these results. We therefore conducted a meta‐analysis to evaluate the effect of IL‐2 for remission maintenance in people with AML. Although Buyse 2011 conducted an individual patient data meta‐analysis to evaluate the effect of IL‐2, they only included the use of IL‐2 monotherapy for remission maintenance in people with AML in first CR. Our review tried to provide more comprehensive evidence for the efficacy of IL‐2 as maintenance therapy by including both people receiving IL‐2 monotherapy and people receiving IL‐2 combination therapy.

Objectives

To evaluate the efficacy and safety of IL‐2 as maintenance therapy for children and adults with AML who have achieved first CR and have not relapsed.

Methods

Criteria for considering studies for this review

Types of studies

We conducted the current systematic review according to the published protocol (Mao 2012). We considered only RCTs. We included both full‐text and abstract publications.

Types of participants

Participants were adults and children diagnosed with AML who had achieved first CR but had not relapsed. We applied no restriction on definitions of AML. All participants achieved first CR before randomisation to maintenance treatment.

Types of interventions

We applied no restriction on treatments patients received in induction and consolidation phases before randomisation to maintenance treatment. We included patients who received allo‐SCT, auto‐SCT, or chemotherapy alone in induction and consolidation phases. For this systematic review, we included the following comparisons:

  • IL‐2 maintenance monotherapy (intervention) versus placebo (control) or best supportive care (control) or maintenance chemotherapy (control);

  • IL‐2 plus maintenance chemotherapy (intervention) versus the same maintenance chemotherapy alone (control).

We excluded studies that compared IL‐2 in treatment group with allogenic or autologous stem cell transplantation in control group.

Types of outcome measures

The primary and secondary outcomes are listed as follows.

Primary outcomes

  • Disease‐free survival: defined as the time interval from the date of remission to leukaemic relapse or death from any cause.

  • Overall survival: defined as the time interval from the date of randomisation or entry into study to death from any cause.

Secondary outcomes

  • Event‐free survival: defined as the time interval from the date of randomisation or entry into study to CR achievement failure, first relapse, or death from any cause.

  • Treatment‐related mortality: defined as the time interval from the date of randomisation or entry into study to death, resulting from non‐progressive disease where induction failures, relapses, and deaths resulting from progressive disease were competing events.

  • Adverse events.

  • Quality of life.

Search methods for identification of studies

Electronic searches

We adapted search strategies as suggested in the Cochrane Handbook for Systematic Reviews of Interventions (Lefebvre 2011). In order to reduce language bias, we imposed no language restrictions.

We searched the following databases of medical literature.

  1. Cochrane Central Register of Controlled Trials (CENTRAL) (Cochrane Library 2015, Issue 8, see Appendix 1 for search strategy).

  2. MEDLINE (1950 to August 2015, see Appendix 2 for search strategy).

  3. EMBASE (1950 to August 2015, see Appendix 3 for search strategy).

  4. LILACS (1982 to August 2015, see Appendix 4 for search strategy).

  5. Chinese BioMedical Literature Database (CBM) (1978 to August 2015, see Appendix 5 for search strategy).

Searching other resources

We searched the following conference proceedings.

  1. American Society of Hematology (from 2000 to 2015).

  2. American Society of Clinical Oncology (from 2000 to 2015).

  3. European Hematology Association (from 2000 to 2015).

  4. European Society of Medical Oncology (from 2000 to 2015).

  5. International Society for Hematology and Stem Cells (available at: http://www.exphem.org/search) (from 1999 to 2015).

We electronically searched the metaRegister of Controlled Trials (available at: http://www.isrctn.com/page/mrct) for ongoing trials (since inception to August 2015, see Appendix 6 for search strategy) and saved the search results of each item for further screening.

We also checked the references of included studies and major reviews for additional studies.

Data collection and analysis

Selection of studies

We performed the selection of studies according to the guidelines in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011a). Two review authors (CM and XF) independently scanned the title and abstract, then investigated all potentially relevant articles as full text by using an eligibility form addressing study design and compliance with inclusion criteria. The forms contained the following questions.

  • Is the study described as randomised?

  • Were the participants diagnosed with AML?

  • Had the participants achieved first CR and not relapsed?

  • Were the participants in the intervention group treated with IL‐2?

Any disagreements were resolved by discussion of two review authors (CM and XF) or by consulting a third review author (JT). We used a PRISMA flow diagram to show numbers of identified citations, excluded studies, and included studies (Moher 2009).

Data extraction and management

We extracted data from the included trials according to guidelines in Chapter 7 of the Cochrane Handbook for Systematic Reviews of Interventions by using a predesigned data extraction form (Higgins 2011a). Two review authors (CM and XF) conducted the data extraction independently. Any disagreements were resolved by discussion of two review authors or by consulting a third review author (JT). We extracted information for the following items.

  • General information: author, title, source, publication date, country, language, and duplicate publications.

  • Quality assessment: sequence generation, allocation concealment, blinding (participants and outcome assessors), incomplete outcome data, selective outcome reporting, and other sources of bias.

  • Study characteristics: trial design, aims, setting and dates, source of participants, inclusion/exclusion criteria, comparability of groups, statistical methods, power calculations, treatment cross‐overs, compliance with assigned treatment, length of follow‐up, and time point of randomisation.

  • Participant characteristics: age, gender, number of participants recruited/allocated/evaluated, and participants lost to follow‐up.

  • Interventions: setting, dose, and duration of IL‐2 treatment, type of additional or comparator chemotherapy, supportive treatment.

  • Outcomes: disease‐free survival, overall survival, event‐free survival, treatment‐related mortality, adverse events, and quality of life.

Assessment of risk of bias in included studies

Two review authors (ZY and JY) independently evaluated the quality of the included studies using The Cochrane Collaboration's tool for assessing risk of bias in randomised trials (Higgins 2011b). Any disagreements were resolved by discussion of two review authors or by consulting a third review author (JT). When necessary, we contacted the authors of included studies for clarification. We assessed the risk of bias according to the following domains.

  • Sequence generation;

  • Allocation concealment;

  • Blinding (participants and outcome assessors);

  • Incomplete outcome data;

  • Selective outcome reporting;

  • Other sources of bias.

We graded each domain as high, low, or unclear risk of bias (Higgins 2011b).

Measures of treatment effect

For time‐to‐event outcomes including disease‐free survival, overall survival, treatment‐related mortality, and event‐free survival, we expressed the effect measure as a hazard ratio (HR) with 95% confidence interval (CI). In trials that did not report HR, we calculated the HR for each endpoint from observed minus expected number of events and variance. If we were unable to extract data for time‐to‐event outcomes, we estimated 'O‐E' and 'V' statistics indirectly using the methods described by Parmar 1998 and Tierney 2007.

For dichotomous data including adverse effects, we expressed the effect measure as a risk ratio (RR) with 95% CI. For continuous outcomes including quality of life, we expressed the effect measure as weighted mean differences with 95% CI.

Unit of analysis issues

The primary analysis was per individual randomised. We collected and analysed a single measurement for each outcome from each participant. We would have included studies with more than two treatment groups and presented the additional treatment arms. We would have acknowledged heterogeneity in the randomisation unit and performed a sensitivity analysis.

Dealing with missing data

Regarding missing data, we planned to make assumptions (for example that the data were missing at random or not missing at random) based on the report of reasons for missing data. If we assumed the data were missing at random, we only analysed the available data (that is ignoring the missing data). If we assumed the data were not missing at random, we input the missing data with replacement values, and treated these as if they were observed (for example last observation carried forward, imputing an assumed outcome such as assuming all were poor outcomes, imputing the mean, and imputing based on predicted values from a regression analysis). We planned to perform sensitivity analysis of the primary outcome on these assumptions in order to detect the impact of assumption modifications on the results.

Assessment of heterogeneity

We examined heterogeneity of treatment effects among trials qualitatively with the Chi2 test at a significance level of 0.1 according to Chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011c). We used the I2 statistic to quantify possible heterogeneity based on the following rough guide (Deeks 2011):

  • 0% to 40%: might not be important;

  • 30% to 60%: may represent moderate heterogeneity;

  • 50% to 90%: may represent substantial heterogeneity;

  • 75% to 100%: considerable heterogeneity.

Assessment of reporting biases

As suggested in Chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions (Sterne 2011), in meta‐analyses involving at least 10 trials, we would have detected potential reporting bias by generating funnel plots and testing reporting bias statistically with a linear regression test. In this systematic review, it was not possible to conduct funnel plots due to lack of data, see Differences between protocol and review.

Data synthesis

We performed statistical analyses according to Chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions with Review Manager (Higgins 2011c; RevMan 2012). We performed meta‐analysis using fixed‐effect model. For time‐to‐event data including disease‐free survival, overall survival, treatment‐related mortality, and event‐free survival, we combined the HRs using the generic inverse‐variance method (Mantel 1959). For dichotomous data like adverse events, we intended to combine RRs using the Mantel‐Haenszel method. For continuous outcomes like quality of life, we intended to combine weighted mean differences using the inverse‐variance method.

As recommended in Chapter 11 of the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2011), we created a 'Summary of findings' table using GRADEpro software (GRADEpro 2008). Our prioritised outcomes in the summary of findings Table for the main comparison table were overall survival, disease‐free survival, event‐free survival, treatment‐related mortality, adverse events, and quality of life.

Subgroup analysis and investigation of heterogeneity

We explored potential sources of heterogeneity through subgroup analyses based on the following factors:

  • age (younger adults age < 60 years versus older adults age ≥ 60 years);

  • different classifications of AML (> 20% blasts versus ≥ 30% blasts);

  • treatment used before the maintenance randomisation (allo‐SCTs versus auto‐SCTs versus chemotherapy alone);

  • duration of IL‐2 treatment (≤ 3 months versus > 3 months).

We had intended to carry out subgroup analysis of children, but data for subgroup of children were not available.

Sensitivity analysis

We performed sensitivity analyses by excluding abstracts, changing to the alternative analysis model (fixed‐effect model versus random‐effects model), omitting studies at high risk of bias, and including all studies reporting disease‐free survival or event‐free survival.

Results

Description of studies

Results of the search

Our search identified 679 citations, of which 22 references were considered as potentially relevant and were subjected to full‐text assessment. After reviewing the full text, we excluded eight trials and included nine studies with 14 citations that fulfilled the defined inclusion criteria. See flow chart (Figure 1) for details of the search results.


Flow diagram of literature search

Flow diagram of literature search

Included studies

Nine trials with 14 publications were eligible, of which six were abstracts (Faber 1997; Kolitz 2014; Petit 2014; Willemze 2011), and eight were full texts (Baer 2008; Blaise 2000; Kolitz 2014; Lange 2011; Liu 2011; Pautas 2010). The number of participants enrolled was 1665. The trials randomised between 24 and 528 participants, comparing IL‐2 with no further treatment. All participants had achieved first CR before randomisation to IL‐2 or no further treatment. Participants were recruited between 1991 to 2008, with a median follow‐up period ranging from to 2.4 to 8.3 years. Funding sources were the National Cancer Institute to the Cancer and Leukemia Group B (Baer 2008; Kolitz 2014), the La Ligue Nationale de Lutte contre le Cancer (Blaise 2000), the National Institutes of Health (Lange 2011), the Acute Leukemia French Association (Pautas 2010), the European Organisation for Research and Treatment of Cancer (Willemze 2011), and the Assistance Publique ‐ Hôpitaux de Paris (Petit 2014). Faber 1997 and Liu 2011 did not report funding sources, and none of the included studies was funded by the drug manufacturer or by an agency with a commercial interest in the results of the studies.

Study design

All studies were parallel‐group RCTs.

Participants

The nine trials included a total of 1665 participants, and the number of participants in each study ranged from 24, in Faber 1997, to 528, in Willemze 2011. Except for one study conducted in China (Liu 2011), all studies took place in Western countries.

In four studies, the participants were no more than 60 years old (Blaise 2000; Lange 2011; Kolitz 2014; Willemze 2011); in study Baer 2008 participants were above 60; in studies Pautas 2010 and Faber 1997 the age of participants ranged from 29 to 75 years old. Three studies included both children and adults: Liu 2011 included participants from 14 to 72 years old; Petit 2014 included participants from 0 to 18 years old; and Lange 2011 included participants from 1 month to 21 years. Data for subgroups of children and adults were not provided separately.

Two studies applied the WHO classification to define AML (greater than 20% blasts) (Kolitz 2014; Petit 2014), five studies applied the the FAB classification to define AML (equal to or greater than 30% blasts) (Baer 2008; Lange 2011; Liu 2011; Pautas 2010; Willemze 2011), while the remaining two studies did not specify the criteria used to define AML (Blaise 2000; Faber 1997).

Before randomisation to maintenance treatment, participants received induction and consolidation therapy, which consisted of different combinations of cytarabine, daunorubicin, etoposide, and idarubicin. In three studies, participants received SCT as consolidation therapy before randomisation: all participants received an auto‐SCT in the Blaise 2000 and Willemze 2011 studies, and some participants received an auto‐SCT or allo‐SCT, while others received only chemotherapy because of no available donor or being unsuitable for SCT in the Kolitz 2014 study. Participants who achieved first CR after induction and consolidation therapy and had not relapsed were randomised to maintenance therapy with or without IL‐2.

Interventions

All participants in the treatment groups were treated with IL‐2 monotherapy. Doses and durations of IL‐2 varied across different trials. Daily doses of IL‐2 ranged from low‐dose sequence of 0.2 x 106 IU/m2 in Faber 1997 for expanding cytotoxic effector cells to higher‐dose bolus of 12 to 15 x 106 IU/m2 in Kolitz 2014 for activating cytotoxic effector cells. The low dose was generally self administered at home (Lange 2011; Pautas 2010), whereas the higher doses were administered in the outpatient setting under observation (Baer 2008). The duration of intervention ranged from 14 days to one year.

All participants in the control group did not receive any maintenance treatment.

We did not identify studies with comparison of IL‐2 versus best supportive care or maintenance chemotherapy and comparison of IL‐2 plus maintenance chemotherapy versus maintenance chemotherapy alone.

Outcomes

The included studies reported the following outcomes.

Disease‐free survival, overall survival, and event‐free survival were calculated from the time of randomisation to maintenance therapy. The difference between disease‐free survival and event‐free survival was the events included. Disease‐free survival included events of first relapse and death from any cause, while event‐free survival included CR achievement failure, first relapse, and death from any cause. Four studies specified that outcomes mentioned above were calculated from the time of randomisation to maintenance therapy (Baer 2008; Kolitz 2014; Pautas 2010; Willemze 2011); Lange 2011 stated that disease‐free survival was calculated from the end of consolidation therapy, and overall survival was calculated from study entry to death. The remaining studies did not provide information about this.

Excluded studies

We excluded eight trials after reviewing the full text: three were not RCTs (Benyunes 1993; Sievers 1998; Stone 2008); two were not studying people with AML in first CR (Hiddemann 1994; Meloni 1997); Ganser 2000 compared low dose with higher dose of IL‐2 and did not fulfil the criteria we mentioned about type of comparisons; one compared histamine dihydrochloride plus IL‐2 with no treatment (Brune 2009); and one was an ongoing trial (Guy 2012). For more details, see also the Characteristics of excluded studies table.

Risk of bias in included studies

We summarised risk of bias of eight studies reporting overall survival or disease‐free survival in this section. We judged all studies to be at unclear risk of bias in terms of overall survival and at high risk of bias in terms of disease‐free survival. We tried to contact the original authors but received no reply. For more details, see also the Characteristics of excluded studies table; Figure 2 and Figure 3.


Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.


Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

All nine studies mentioned randomisation, however none of them described the method of sequence generation in detail. Allocation concealment was adequate in two studies (Lange 2011; Willemze 2011), which reported that they performed randomisation centrally. The method of allocation concealment was unclear in the remaining six studies. We thus judged all studies as at unclear risk of selection bias.

Blinding

Five studies were open‐label RCTs (Baer 2008; Kolitz 2014; Lange 2011; Pautas 2010; Petit 2014), and the remaining three studies did not mention blinding. However, the results of disease‐free survival and overall survival were determined mainly by biological, objective effect of treatments and performance of outcome assessors, and were unlikely to be affected by participants' or personnel's knowledge of assignment status. We thus judged the risk of performance bias as low.

We judged risk of detection bias for overall survival as low, because death was an objective hard outcome and unlikely to be affected by outcome assessors' knowledge of treatment status. We judged the risk of detection bias for disease‐free survival as high, because the results of disease‐free survival consisted of death and AML relapse, which involved subjective judgements and were vulnerable to the performance of assessors who knew the assignment status.

Incomplete outcome data

All studies conducted analyses on an intention‐to‐treat basis except for Faber 1997, which did not mention the use of an intention‐to‐treat analysis or participant withdrawals. We thus judged Faber 1997 as at unclear risk of attrition bias and the remaining seven trials as at low risk of attrition bias.

Selective reporting

We found no protocol or registration information for the studies Blaise 2000 and Faber 1997, which reported disease‐free survival and overall survival without other secondary outcomes. Risk of reporting bias for these two studies was unclear. The remaining six studies reported all prespecified outcomes, and were thus judged as at low risk of reporting bias.

Other potential sources of bias

Faber 1997 did not provide detailed information on date of randomisation and participant characteristics, and was thus judged as at unclear risk of bias, as participant characteristics between the IL‐2 group and the control group could be imbalanced. The remaining studies all reported that randomisation took place at the end of consolidation therapy or on the date of transplantation.

Effects of interventions

See: Summary of findings for the main comparison

Primary outcomes

1. Disease‐free survival

The median disease‐free survival with IL‐2 maintenance therapy ranged from 7.5 months, in Baer 2008, to 15 months, in Blaise 2000 (median: 12.1 months), while the median disease‐free survival with no treatment ranged from 5.8 months, in Baer 2008, to 15 months, in Blaise 2000 (median: 9.5 months). We conducted meta‐analysis based on six studies that provided suitable data (Baer 2008; Blaise 2000; Kolitz 2014; Lange 2011; Petit 2014; Willemze 2011), involving 1426 participants (Analysis 1.1, Figure 4). The pooled HR was 0.95 (95% CI 0.86 to 1.06, P = 0.37; quality of evidence: low), indicating there was no evidence of difference in disease‐free survival between IL‐2 and no treatment. We observed no statistical heterogeneity among the studies (P = 0.78, I2 = 0%). We did not include Faber 1997 in the meta‐analysis because the study did not report HRs or related information. In this study, no significant difference was observed between the IL‐2 group and the control group, and the median disease‐free survival in the IL‐2 group and control group was 12.1 months and 9.5 months, respectively.


Forest plot of comparison: 1 IL‐2 vs. Control, outcome: 1.1 Disease‐free survival.

Forest plot of comparison: 1 IL‐2 vs. Control, outcome: 1.1 Disease‐free survival.

2. Overall survival

The median overall survival with IL‐2 maintenance therapy ranged from 9.3 months, in Pautas 2010, to 17.2 months, in Faber 1997 (median: 14 months), while the median overall survival with no treatment ranged from 13 months, in Faber 1997, to 31.4 months, in Pautas 2010 (median: 16.4 months). We conducted meta‐analysis based on five studies that provided suitable data (Baer 2008; Kolitz 2014; Lange 2011; Pautas 2010; Willemze 2011), involving 1355 participants (Analysis 1.2, Figure 5). The pooled HR was 1.05 (95% CI 0.95 to 1.16, P = 0.35; quality of evidence: moderate), indicating there was no evidence of difference in overall survival between IL‐2 and no treatment. We observed no significant statistical heterogeneity among the studies (P = 0.53, I2 = 0%). We did not combine results of the remaining two studies because they did not report HRs or related information (Blaise 2000; Faber 1997). In these two studies, no significant difference was observed between the IL‐2 group and the control group: Blaise 2000 reported that the estimate of 5‐year overall survival rate for participants in the IL‐2 group and control group was 38% (24 to 53) and 47% (32 to 62), respectively; Faber 1997 reported that the median overall survival in the IL‐2 group and control group was 17.2 (8.9 to 37.3) months and 13 (6.5 to 35.2) months, respectively.


Forest plot of comparison: 1 IL‐2 vs. Control, outcome: 1.2 Overall survival.

Forest plot of comparison: 1 IL‐2 vs. Control, outcome: 1.2 Overall survival.

Secondary outcomes

1. Event‐free survival

One study reported data on event‐free survival (Pautas 2010). The median event‐free survival of the IL‐2 group and control group was 10.3 months and 9.3 months, respectively. HR estimated from the survival curve was 1.02 (95% CI 0.79 to 1.32, P = 0.88; quality of evidence: low), indicating that there was no significant difference in event‐free survival between the IL‐2 group and control group.

2. Adverse events

Four studies reported data on adverse events. One of them compared IL‐2 with no treatment in terms of adverse events (Willemze 2011), while the other three reported adverse events in the IL‐2 group only (Baer 2008; Kolitz 2014; Pautas 2010). Willemze 2011 observed a much higher incidence of hypersensitivity, fatigue, rigor/chills, and arthralgia/myalgia in the IL‐2 group than in the control group. The corresponding RRs for hypersensitivity, fatigue, rigor/chills, and arthralgia/myalgia were (RR 17.13; 95% CI 0.99 to 295.26, P = 0.05; quality of evidence: very low), (RR 7.05; 95% CI 2.13 to 23.36, P = 0.001; quality of evidence: very low), (RR 33.25; 95% CI 2.01 to 551.36, P = 0.01; quality of evidence: very low), and (RR 19.14; 95% CI 1.12 to 327.24, P = 0.04; quality of evidence: very low), respectively (Analysis 1.3).

The other three studies reported adverse events in participants treated with IL‐2, which are shown in Table 1. The majority of adverse events were thrombocytopenia, neutropenia, malaise and fatigue, and infection and fever. No mortality due to adverse events was reported.

Open in table viewer
Table 1. Adverse events

Adverse events

Thrombocytopenia

Neutropenia

Malaise/fatigue

Infection/fever

Baer 2008

65%

64%

15%

24%

Kolitz 2014

24%

20%

12%

5%

Pautas 2010

11%

3%

5%

57%

3. Treatment‐related mortality and quality of life

None of the included studies provided data on treatment‐related mortality or quality of life.

Subgroup analyses

We carried out subgroup analyses according to age of adults, classifications of AML, treatment used before the maintenance randomisation, and duration of IL‐2 treatment regarding disease‐free survival (Analysis 1.4; Analysis 1.5; Analysis 1.6; Analysis 1.7) and overall survival (Analysis 1.8; Analysis 1.9; Analysis 1.10; Analysis 1.11). We had intended to carry out subgroup analysis in paediatric participants (age < 15 years versus ≥ 15 years). Two studies included paediatric participants (Lange 2011; Petit 2014), but did not provide data for subgroups, so we were unable to conduct subgroup analysis based on children's age.

Subgroup analyses of younger adults (age < 60) versus older adults (age ≥ 60) found no significant differences in disease‐free survival (P = 0.76, I2 = 0%, Analysis 1.4) or overall survival (P = 0.25, I2 = 25.4%, Analysis 1.8). Based on data available on classifications ( > 20% blasts versus ≥ 30% blasts), we observed no significant subgroup differences in disease‐free survival (P = 0.37, I2 = 0%, Analysis 1.5) or overall survival (P = 0.13, I2 = 55.7%, Analysis 1.9) between different AML classifications. Analyses of chemotherapy alone and auto‐SCTs did not reveal significant subgroup differences in disease‐free survival (P = 0.86, I2 = 0%, Analysis 1.6) or overall survival (P = 0.38, I2 = 0%, Analysis 1.11). Analyses of IL‐2 treatment duration of ≤ 3 months and > 3 months showed no significant difference with respect to disease‐free survival (P = 0.98, I2 = 0%, Analysis 1.7) or overall survival (P = 0.46, I2 = 0%, Analysis 1.10). The results are shown in Table 2 and Table 3.

Open in table viewer
Table 2. Subgroup analyses for disease‐free survival

Groups

No. of studies

HR [95% CI]

Test for subgroup differences

Age

< 60 years

3

0.92 [0.78, 1.08]

P = 0.76, I2 = 0%

≥ 60 years

1

0.95 [0.80, 1.13]

AML classification

> 20% blasts

2

0.86 [0.68, 1.09]

P = 0.37, I2 = 0%

≥ 30% blasts

3

0.98 [0.86, 1.10]

Treatment before IL‐2

Chemotherapy alone

3

0.98 [0.86, 1.12]

P = 0.86, I2 = 0%

Autologous stem cell transplantation

2

0.96 [0.80, 1.15]

IL‐2 duration

≤ 3 months

4

0.95 [0.84, 1.08]

P = 0.98, I2 = 0%

> 3 months

2

0.95 [0.80, 1.14]

CI: confidence interval
HR: hazard ratio
IL: interleukin‐2

Open in table viewer
Table 3. Subgroup analyses for overall survival

Groups

No. of studies

HR [95% CI]

Test for subgroup differences

Age

< 60 years

2

0.93 [0.77, 1.11]

P = 0.25, I2 = 25.4%

≥ 60 years

1

1.11 [0.87, 1.40]

AML classification

> 20% blasts

1

0.88 [0.68, 1.13]

P = 0.13, I2 = 55.7%

≥ 30% blasts

4

1.09 [0.97, 1.22]

Treatment before IL‐2

Chemotherapy alone

3

1.11 [0.98, 1.26]

P = 0.38, I2 = 0%

Autologous stem cell transplants

1

0.98 [0.76, 1.26]

IL‐2 duration

≤ 3 months

3

1.00 [0.85, 1.18]

P = 0.46, I2 = 0%

> 3 months

2

1.08 [0.95, 1.24]

CI: confidence interval
HR: hazard ratio
IL: interleukin‐2

Sensitivity analyses

We conducted sensitivity analyses by excluding abstracts and changing to random‐effects model (Analysis 1.12; Analysis 1.13; Analysis 1.14; Analysis 1.15). We did not conduct sensitivity analysis based on risk of bias, because we judged all studies as at unclear risk of bias in terms of overall survival and all studies as at high risk of bias in terms of disease‐free survival. Excluding two abstract publications (Petit 2014; Willemze 2011), meta‐analyses of disease‐free survival (Baer 2008; Blaise 2000; Kolitz 2014; Lange 2011), and of overall survival (Baer 2008; Kolitz 2014; Lange 2011; Pautas 2010), did not reveal significant differences compared with full analysis; HRs were 0.95 (95% CI 0.84 to 1.08, P = 0.47) and 1.06 (95% CI 0.95 to 1.19, P = 0.28), respectively. Changing to random‐effects model, results of meta‐analyses of disease‐free survival (Baer 2008; Blaise 2000; Kolitz 2014; Lange 2011; Petit 2014; Willemze 2011), and of overall survival (Baer 2008; Kolitz 2014; Lange 2011; Pautas 2010; Willemze 2011), were not significantly different from the results using fixed‐effect model. We conducted sensitivity analysis by including all studies reporting data on disease‐free survival or event‐free survival (Analysis 1.16). HR of the full analysis of seven studies was 0.96 (95% CI 0.87 to 1.06, P = 0.44) (Baer 2008; Blaise 2000; Kolitz 2014; Lange 2011; Pautas 2010; Petit 2014; Willemze 2011), and there was no significant difference from result of the meta‐analysis including only disease‐free survival. The results are shown in Table 4.

Open in table viewer
Table 4. Sensitivity analyses

Outcomes

Sensitivity analyses

No. of studies

HR [95% CI]

Disease‐free survival

Excluding 2 abstracts

4

0.95 [0.84, 1.08]

Random‐effects model

6

0.95 [0.86, 1.06]

Overall survival

Excluding 2 abstracts

4

1.06 [0.95, 1.19]

Random‐effects model

5

1.05 [0.95, 1.16]

Disease‐free survival and event‐free survival

Excluding event‐free survival

6

0.95 [0.86, 1.06]

Including event‐free survival

7

0.96 [0.87, 1.06]

CI: confidence interval
HR: hazard ratio

Discussion

Summary of main results

We included nine RCTs enrolling a total of 1665 participants, and eight reported outcomes of interest. We summarised the main findings of this review in the summary of findings Table for the main comparison.

  • Disease‐free survival: Seven studies reported disease‐free survival, six of which with 1426 participants we included in the meta‐analysis. There was no evidence for a difference between IL‐2 maintenance therapy and no treatment with respect to disease‐free survival of people with AML in first CR.

  • Overall survival: Seven studies reported overall survival, five of which with 1355 participants we included in the meta‐analysis. There was no evidence for a difference between IL‐2 maintenance therapy and no treatment regarding overall survival of people with AML in first CR.

  • Event‐free survival: There was no evidence for a difference between IL‐2 maintenance therapy and no treatment regarding event‐free survival of people with AML in first CR based on one single study of 161 participants.

  • Adverse events: Weak evidence indicated that adverse events (including hypersensitivity, fatigue, rigor/chills, and arthralgia/myalgia) seemed to be more frequent in participants treated with IL‐2 than in those who received no treatment. No mortality due to adverse events was reported.

  • Treatment‐related mortality and quality of life: No study reported treatment‐related mortality or quality of life.

Disease‐free survival, overall survival, and event‐free survival were calculated from the time of randomisation to maintenance therapy. The difference between disease‐free survival and event‐free survival was the events included. Disease‐free survival included events of first relapse and death from any cause, while event‐free survival included CR achievement failure, first relapse, and death from any cause. We conducted sensitivity analysis by including all studies reporting disease‐free survival or event‐free survival, and the result revealed no significant difference from that of meta‐analysis including only disease‐free survival.

Subgroup analyses did not identify any significant subgroup differences among participants of different subgroups based on age, classification of AML, treatment used before the maintenance randomisation, and IL‐2 duration. Sensitivity analyses did not reveal significant difference in main results.

Overall completeness and applicability of evidence

The completeness of the evidence summarised in this review was unsatisfactory. As data on secondary outcomes were insufficient, we were unable to carry out meta‐analysis for event‐free survival, treatment‐related mortality, adverse events, and quality of life. According to the inclusion criteria of eligible studies, the results of the present review are applicable to people with AML in first CR regardless of age, induction therapy, and IL‐2 treatment duration. However, existing results were calculated from Western studies, so caution is required in applying the evidence to patients of other origins, as survival prospects varied in people with AML in different ethnic groups (Aplenc 2006).

Quality of the evidence

All included studies were RCTs. Allocation generation and concealment were unclear in the included trials, which might introduce selection bias for all trials. One study reporting in abstract did not provide participant characteristics (Faber 1997), and it is possible there is an imbalance between the IL‐2 group and control group. We could not assess the potential risk of bias in detail for this trial. None of the trials reported blinding of participants and personnel or outcome assessors, which could lead to detection bias with regard to disease‐free survival and event‐free survival. We thus judged all studies as at low risk of bias in terms of overall survival and as at high risk of bias in terms of disease‐free survival. Fortunately, heterogeneity tests suggested that results were highly consistent across all included studies, and sensitivity analyses did not change the main results considerably.

Using the GRADE scoring system, we graded the quality of evidence for outcomes as moderate to very low. We judged the quality of the evidence as low for disease‐free survival and event‐free survival due to high risk of detection bias and potential selection bias. We considered the evidence for overall survival as moderate due to the potential selection bias. The quality of evidence for adverse events was very low due to study limitations and wide confidence intervals. All RCTs evaluated the effectiveness of IL‐2 using direct comparisons between IL‐2 and no treatment. In addition, all included studies measured outcomes directly, without using surrogate outcomes. Thus, there was no indication of indirectness of the outcomes, as all RCTs evaluated the effectiveness of IL‐2. We could not assess the quality of evidence for treatment‐related mortality or quality of life due to limited reported information.

Potential biases in the review process

We attempted to minimise bias in the review process. We conducted a comprehensive search without language restrictions to identify all relevant studies. We also searched trial registries and checked the citations of included trials and relevant reviews. Two review authors independently carried out study selection, assessment of risk of bias, and data extraction. However, potential biases exist. Firstly, two studies did not provide HR values or survival curve, so we did not include them in our meta‐analyses (Faber 1997 for disease‐free survival and overall survival, Blaise 2000 for overall survival). However, results of these two studies were consistent with the pooled estimates. Secondly, some studies provided survival curves or survival rates with P values for logrank tests without HR values, so we had to estimate HRs from the information reported, which might introduce bias. Thirdly, Liu 2011 did not report any relevant outcomes for this review. It has been suggested that some authors may choose not to report negative outcomes so their studies are more likely to be published. Liu 2011 aimed to investigate the influence of IL‐2 on immune function and mainly focused on the changes in lymphocyte subgroups. We have contacted the authors of this study and confirmed that other outcomes were not evaluated, therefore the reason why other outcomes were not reported was not because the author obtained negative results that they chose not to report, and this study would not raise the risk of reporting bias. Lastly, as we included fewer than 10 studies, we could not use pre‐planned funnel plots to detect publication bias. Thus, we cannot exclude the possibility of publication bias in the present review.

There is one ongoing study (Guy 2012), an open‐label RCT starting in March 2005. People with AML up to 18 years old not being allogeneic transplanted in CR after induction and consolidation chemotherapy were randomised to receive IL‐2 treatment for five days each month during one year or no treatment. The primary outcome was event‐free survival. The study was supposed to accrue 580 participants over seven years and complete in 2012. We contacted the researchers for updated status of the study but received no reply. We updated the search for ongoing studies in August 2015 and identified no more eligible studies. However, this does not imply that there is no need for future research on IL‐2 in people with AML. There are ongoing trials of IL‐2 in combination with other treatments in people with AML in different phases, which are not eligible for our review. For the knowledge gap that future research needs to fulfil, see Implications for research.

Agreements and disagreements with other studies or reviews

Our findings are generally consistent with the published review Buyse 2011 and other related studies.

Buyse 2011 evaluated the effect of IL‐2 monotherapy compared with no treatment in people with AML who had achieved first CR. Meta‐analyses of individual participant data from six RCTs were performed with respect to endpoints leukaemia‐free survival and overall survival, and reached the conclusion that IL‐2 was not significantly different from no treatment as a remission maintenance therapy for people with AML in first CR (Baer 2008; Blaise 2000; Kolitz 2014; Lange 2011; Pautas 2010; Willemze 2011).

An RCT that assessed the effect of IL‐2 in people with relapsed or refractory AML and in people with AML in second CR observed no significant benefit in disease‐free survival or overall survival (Meloni 1997). An RCT comparing a higher dose (9 x 106 IU/m2) with a low dose (0.9 x 106 IU/m2) of IL‐2 in people with AML after consolidation therapy observed no significant difference in disease‐free survival or overall survival (Ganser 2000), which is in accordance with our finding that IL‐2 efficacy was similar across studies with different doses of IL‐2.

A recent RCT evaluated the long‐term effect of post‐consolidation immunotherapy with IL‐2 and histamine dihydrochloride (HDC) in people with AML in first or subsequent CR. There was a significant benefit in disease‐free survival in participants treated with combination therapy with IL‐2 and HDC in three and six years' follow‐up (Brune 2006; Brune 2009). Preclinical studies have demonstrated that HDC improves the antileukaemic function of IL‐2 by stimulating NK cell activity, thus the combination of IL‐2 and HDC was more effective than either of them alone (Brune 1996; Romero 2009).

Flow diagram of literature search
Figures and Tables -
Figure 1

Flow diagram of literature search

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Figures and Tables -
Figure 2

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Figures and Tables -
Figure 3

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Forest plot of comparison: 1 IL‐2 vs. Control, outcome: 1.1 Disease‐free survival.
Figures and Tables -
Figure 4

Forest plot of comparison: 1 IL‐2 vs. Control, outcome: 1.1 Disease‐free survival.

Forest plot of comparison: 1 IL‐2 vs. Control, outcome: 1.2 Overall survival.
Figures and Tables -
Figure 5

Forest plot of comparison: 1 IL‐2 vs. Control, outcome: 1.2 Overall survival.

Comparison 1 IL‐2 versus control, Outcome 1 Disease‐free survival.
Figures and Tables -
Analysis 1.1

Comparison 1 IL‐2 versus control, Outcome 1 Disease‐free survival.

Comparison 1 IL‐2 versus control, Outcome 2 Overall survival.
Figures and Tables -
Analysis 1.2

Comparison 1 IL‐2 versus control, Outcome 2 Overall survival.

Comparison 1 IL‐2 versus control, Outcome 3 Adverse events.
Figures and Tables -
Analysis 1.3

Comparison 1 IL‐2 versus control, Outcome 3 Adverse events.

Comparison 1 IL‐2 versus control, Outcome 4 Subgroup analysis of DFS: age<60 versus age ≥60.
Figures and Tables -
Analysis 1.4

Comparison 1 IL‐2 versus control, Outcome 4 Subgroup analysis of DFS: age<60 versus age ≥60.

Comparison 1 IL‐2 versus control, Outcome 5 Subgroup analysis of DFS: >20% blasts versus ≥30% blasts.
Figures and Tables -
Analysis 1.5

Comparison 1 IL‐2 versus control, Outcome 5 Subgroup analysis of DFS: >20% blasts versus ≥30% blasts.

Comparison 1 IL‐2 versus control, Outcome 6 Subgroup analysis of DFS: chemotherapy alone versus autologous stem cell transplantation.
Figures and Tables -
Analysis 1.6

Comparison 1 IL‐2 versus control, Outcome 6 Subgroup analysis of DFS: chemotherapy alone versus autologous stem cell transplantation.

Comparison 1 IL‐2 versus control, Outcome 7 Subgroup analysis of DFS: IL‐duration ≤3months versus >3 months.
Figures and Tables -
Analysis 1.7

Comparison 1 IL‐2 versus control, Outcome 7 Subgroup analysis of DFS: IL‐duration ≤3months versus >3 months.

Comparison 1 IL‐2 versus control, Outcome 8 Subgroup analysis of OS: age<60 versus age ≥60.
Figures and Tables -
Analysis 1.8

Comparison 1 IL‐2 versus control, Outcome 8 Subgroup analysis of OS: age<60 versus age ≥60.

Comparison 1 IL‐2 versus control, Outcome 9 Subgroup analysis of OS: >20% blasts versus ≥30% blasts.
Figures and Tables -
Analysis 1.9

Comparison 1 IL‐2 versus control, Outcome 9 Subgroup analysis of OS: >20% blasts versus ≥30% blasts.

Comparison 1 IL‐2 versus control, Outcome 10 Subgroup analysis of OS: IL‐duration ≤3months versus 12 months.
Figures and Tables -
Analysis 1.10

Comparison 1 IL‐2 versus control, Outcome 10 Subgroup analysis of OS: IL‐duration ≤3months versus 12 months.

Comparison 1 IL‐2 versus control, Outcome 11 Subgroup analysis of OS: chemotherapy alone versus autologous stem cell transplantation.
Figures and Tables -
Analysis 1.11

Comparison 1 IL‐2 versus control, Outcome 11 Subgroup analysis of OS: chemotherapy alone versus autologous stem cell transplantation.

Comparison 1 IL‐2 versus control, Outcome 12 Sensitivity analysis of DFS (random‐effects model).
Figures and Tables -
Analysis 1.12

Comparison 1 IL‐2 versus control, Outcome 12 Sensitivity analysis of DFS (random‐effects model).

Comparison 1 IL‐2 versus control, Outcome 13 Sensitivity analysis of DFS (excluding abstracts).
Figures and Tables -
Analysis 1.13

Comparison 1 IL‐2 versus control, Outcome 13 Sensitivity analysis of DFS (excluding abstracts).

Comparison 1 IL‐2 versus control, Outcome 14 Sensitivity analysis of OS (excluding abstracts).
Figures and Tables -
Analysis 1.14

Comparison 1 IL‐2 versus control, Outcome 14 Sensitivity analysis of OS (excluding abstracts).

Comparison 1 IL‐2 versus control, Outcome 15 Sensitivity analysis of OS (random‐effects model).
Figures and Tables -
Analysis 1.15

Comparison 1 IL‐2 versus control, Outcome 15 Sensitivity analysis of OS (random‐effects model).

Comparison 1 IL‐2 versus control, Outcome 16 Sensitivity analysis including DFS and EFS.
Figures and Tables -
Analysis 1.16

Comparison 1 IL‐2 versus control, Outcome 16 Sensitivity analysis including DFS and EFS.

IL‐2 compared with no treatment for people with AML in first complete remission

Patient or population: People with AML in first complete remission

Settings: Maintenance therapy

Intervention: IL‐2

Comparison: No treatment

Outcomes

Illustrative comparative risks[1]

Relative effect
(95% CI)

No of Participants
(studies)

Quality of the evidence
(GRADE)

Comments

Assumed risk

Corresponding risk

No treatment

IL‐2

Relapses/death (instead of disease‐free survival)[2]

Follow‐up at median 5 years

590 per 1000

571 per 1000

(535 to 611)

HR 0.95 (0.86 to 1.06)

1426
(6 studies)

⊕⊕⊝⊝[3]
Low

Mortality (instead of overall survival)[2]

Follow‐up at median 5 years

480 per 1000

497 per 1000

(463 to 532)

HR 1.05 (0.95 to 1.16)

1355
(5 studies)

⊕⊕⊕⊝[4]
Moderate

Death/progress (instead of event‐free survival)[2]

Follow‐up at 5 years

280 per 1000

285 per 1000

(229 to 352)

HR 1.02 (0.79 to 1.32)

161

(1 study)

⊕⊕⊝⊝[3]
Low

Treatment‐related mortality

See comments

See comments

Not estimable

0

(0)

See comments

None of the studies reported treatment‐related mortality

Adverse events: hypersensitivity

Follow‐up at median 6 years

See comments

See comments

RR 17.13 (0.99 to 295.26)

528

(1 study)

⊕⊝⊝⊝[3,5]
Very low

Illustrative comparative risks are not estimable as the assumed risk is zero

Adverse events: fatigue

Follow‐up at median 6 years

10 per 1000

71 per 1000

(21 to 233)

RR 7.05 (2.13 to 23.36)

528

(1 study)

⊕⊝⊝⊝[3,5]
Very low

Adverse events: rigor/chills

Follow‐up at median 6 years

See comments

See comments

RR 33.25 (2.01 to 551.36)

528

(1 study)

⊕⊝⊝⊝[3,5]
Very low

Illustrative comparative risks are not estimable as the assumed risk is zero

Adverse events: arthralgia/myalgia

Follow‐up at median 6 years

See comments

See comments

RR 19.14 (1.12 to 327.24)

528

(1 study)

⊕⊝⊝⊝[3,5]
Very low

Illustrative comparative risks are not estimable as the assumed risk is zero

Quality of life

See comments

See comments

Not estimable

0

(0)

See comments

None of the studies reported quality of life

CI: confidence interval; HR: hazard ratio; IL‐2: interleukin‐2; RR: risk ratio

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

1. The basis for the assumed risk is the median control group risk across studies (The absolute effects come directly from the included study if only one study was included). The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). The HRs are first converted to RRs and the corresponding risks are then calculated from the RRs. Corresponding intervention risk, per 1000 people = 1000 × assumed control risk× RR.

2. Absolute effects were estimated from the HRs, and both are reported in the same row.

3. Downgrading two points for study limitations (high risk of detection bias due to lack of blinding; sequence generation and allocation concealment in some studies were not reported, leading to potential selection bias).

4. Downgrading one point for study limitations (sequence generation and allocation concealment in some studies were not reported, leading to potential selection bias).

5. Downgrading one point for imprecision (wide confidence interval).

Figures and Tables -
Table 1. Adverse events

Adverse events

Thrombocytopenia

Neutropenia

Malaise/fatigue

Infection/fever

Baer 2008

65%

64%

15%

24%

Kolitz 2014

24%

20%

12%

5%

Pautas 2010

11%

3%

5%

57%

Figures and Tables -
Table 1. Adverse events
Table 2. Subgroup analyses for disease‐free survival

Groups

No. of studies

HR [95% CI]

Test for subgroup differences

Age

< 60 years

3

0.92 [0.78, 1.08]

P = 0.76, I2 = 0%

≥ 60 years

1

0.95 [0.80, 1.13]

AML classification

> 20% blasts

2

0.86 [0.68, 1.09]

P = 0.37, I2 = 0%

≥ 30% blasts

3

0.98 [0.86, 1.10]

Treatment before IL‐2

Chemotherapy alone

3

0.98 [0.86, 1.12]

P = 0.86, I2 = 0%

Autologous stem cell transplantation

2

0.96 [0.80, 1.15]

IL‐2 duration

≤ 3 months

4

0.95 [0.84, 1.08]

P = 0.98, I2 = 0%

> 3 months

2

0.95 [0.80, 1.14]

CI: confidence interval
HR: hazard ratio
IL: interleukin‐2

Figures and Tables -
Table 2. Subgroup analyses for disease‐free survival
Table 3. Subgroup analyses for overall survival

Groups

No. of studies

HR [95% CI]

Test for subgroup differences

Age

< 60 years

2

0.93 [0.77, 1.11]

P = 0.25, I2 = 25.4%

≥ 60 years

1

1.11 [0.87, 1.40]

AML classification

> 20% blasts

1

0.88 [0.68, 1.13]

P = 0.13, I2 = 55.7%

≥ 30% blasts

4

1.09 [0.97, 1.22]

Treatment before IL‐2

Chemotherapy alone

3

1.11 [0.98, 1.26]

P = 0.38, I2 = 0%

Autologous stem cell transplants

1

0.98 [0.76, 1.26]

IL‐2 duration

≤ 3 months

3

1.00 [0.85, 1.18]

P = 0.46, I2 = 0%

> 3 months

2

1.08 [0.95, 1.24]

CI: confidence interval
HR: hazard ratio
IL: interleukin‐2

Figures and Tables -
Table 3. Subgroup analyses for overall survival
Table 4. Sensitivity analyses

Outcomes

Sensitivity analyses

No. of studies

HR [95% CI]

Disease‐free survival

Excluding 2 abstracts

4

0.95 [0.84, 1.08]

Random‐effects model

6

0.95 [0.86, 1.06]

Overall survival

Excluding 2 abstracts

4

1.06 [0.95, 1.19]

Random‐effects model

5

1.05 [0.95, 1.16]

Disease‐free survival and event‐free survival

Excluding event‐free survival

6

0.95 [0.86, 1.06]

Including event‐free survival

7

0.96 [0.87, 1.06]

CI: confidence interval
HR: hazard ratio

Figures and Tables -
Table 4. Sensitivity analyses
Comparison 1. IL‐2 versus control

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

1 Disease‐free survival Show forest plot

6

1426

Hazard Ratio (Fixed, 95% CI)

0.95 [0.86, 1.06]

2 Overall survival Show forest plot

5

1355

Hazard Ratio (Fixed, 95% CI)

1.05 [0.95, 1.16]

3 Adverse events Show forest plot

1

Risk Ratio (M‐H, Fixed, 95% CI)

Subtotals only

3.1 Hypersensitivity

1

528

Risk Ratio (M‐H, Fixed, 95% CI)

17.13 [0.99, 295.26]

3.2 Fatigue

1

528

Risk Ratio (M‐H, Fixed, 95% CI)

7.05 [2.13, 23.36]

3.3 Rigor/Chills

1

528

Risk Ratio (M‐H, Fixed, 95% CI)

33.25 [2.01, 551.36]

3.4 Arthralgia/Myalgia

1

528

Risk Ratio (M‐H, Fixed, 95% CI)

19.14 [1.12, 327.24]

4 Subgroup analysis of DFS: age<60 versus age ≥60 Show forest plot

4

983

Hazard Ratio (Fixed, 95% CI)

0.93 [0.83, 1.05]

4.1 Younger adults age<60

3

820

Hazard Ratio (Fixed, 95% CI)

0.92 [0.78, 1.08]

4.2 Older adults age ≥ 60

1

163

Hazard Ratio (Fixed, 95% CI)

0.95 [0.80, 1.13]

5 Subgroup analysis of DFS: >20% blasts versus ≥30% blasts Show forest plot

5

1348

Hazard Ratio (Fixed, 95% CI)

0.95 [0.85, 1.06]

5.1 >20% blasts

2

368

Hazard Ratio (Fixed, 95% CI)

0.86 [0.68, 1.09]

5.2 ≥30% blasts

3

980

Hazard Ratio (Fixed, 95% CI)

0.98 [0.86, 1.10]

6 Subgroup analysis of DFS: chemotherapy alone versus autologous stem cell transplantation Show forest plot

5

1212

Hazard Ratio (Fixed, 95% CI)

0.97 [0.87, 1.08]

6.1 Chemotherapy alone

3

606

Hazard Ratio (Fixed, 95% CI)

0.98 [0.86, 1.12]

6.2 Autologous stem cell transplants

2

606

Hazard Ratio (Fixed, 95% CI)

0.96 [0.80, 1.15]

7 Subgroup analysis of DFS: IL‐duration ≤3months versus >3 months Show forest plot

6

1426

Hazard Ratio (Fixed, 95% CI)

0.95 [0.86, 1.06]

7.1 IL‐2 duration ≤3 months

4

744

Hazard Ratio (Fixed, 95% CI)

0.95 [0.84, 1.08]

7.2 IL‐2 duration >3 months

2

682

Hazard Ratio (Fixed, 95% CI)

0.95 [0.80, 1.14]

8 Subgroup analysis of OS: age<60 versus age ≥60 Show forest plot

3

905

Hazard Ratio (Fixed, 95% CI)

0.99 [0.86, 1.14]

8.1 Younger adults age <60

2

742

Hazard Ratio (Fixed, 95% CI)

0.93 [0.77, 1.11]

8.2 Older adults age ≥ 60

1

163

Hazard Ratio (Fixed, 95% CI)

1.11 [0.87, 1.40]

9 Subgroup analysis of OS: >20% blasts versus ≥30% blasts Show forest plot

5

1355

Hazard Ratio (Fixed, 95% CI)

1.05 [0.95, 1.16]

9.1 >20% blasts

1

214

Hazard Ratio (Fixed, 95% CI)

0.88 [0.68, 1.13]

9.2 ≥30% blasts

4

1141

Hazard Ratio (Fixed, 95% CI)

1.09 [0.97, 1.22]

10 Subgroup analysis of OS: IL‐duration ≤3months versus 12 months Show forest plot

5

1355

Hazard Ratio (Fixed, 95% CI)

1.05 [0.95, 1.16]

10.1 IL‐duration ≤3months

3

666

Hazard Ratio (Fixed, 95% CI)

1.00 [0.85, 1.18]

10.2 IL‐duration 12 months

2

689

Hazard Ratio (Fixed, 95% CI)

1.08 [0.95, 1.24]

11 Subgroup analysis of OS: chemotherapy alone versus autologous stem cell transplantation Show forest plot

4

1141

Hazard Ratio (Fixed, 95% CI)

1.09 [0.97, 1.22]

11.1 Chemotherapy alone

3

613

Hazard Ratio (Fixed, 95% CI)

1.11 [0.98, 1.26]

11.2 Autologous stem cell transplants

1

528

Hazard Ratio (Fixed, 95% CI)

0.98 [0.76, 1.26]

12 Sensitivity analysis of DFS (random‐effects model) Show forest plot

6

1426

Hazard Ratio (Random, 95% CI)

0.95 [0.86, 1.06]

13 Sensitivity analysis of DFS (excluding abstracts) Show forest plot

4

744

Hazard Ratio (Fixed, 95% CI)

0.95 [0.84, 1.08]

14 Sensitivity analysis of OS (excluding abstracts) Show forest plot

4

821

Hazard Ratio (Fixed, 95% CI)

1.06 [0.95, 1.19]

15 Sensitivity analysis of OS (random‐effects model) Show forest plot

5

1355

Hazard Ratio (Random, 95% CI)

1.05 [0.95, 1.16]

16 Sensitivity analysis including DFS and EFS Show forest plot

7

1587

Hazard Ratio (Fixed, 95% CI)

0.96 [0.87, 1.06]

Figures and Tables -
Comparison 1. IL‐2 versus control