Key messages

• Compared to standard dental screw implants placed after building up the jawbone, longer screws placed in the cheekbone were less likely to fail and allowed working replacement teeth to be fitted much sooner, but led to more complications.

• Drilling the holes that receive the implant screws can be done with a standard rotary (spinning) tool or a vibrating tool that 'shaves' the bone (piezoelectric surgery). There may be little or no difference between the two methods in terms of how durable the implants or replacement teeth are, or the risk of complications, but rotary drills are probably faster.

• Few studies have looked at these issues. We need more and bigger studies to have more confidence about the benefits and risks.

What happens when people have no or few natural teeth?

When teeth are lost, the surrounding bone can gradually shrink because it is no longer stimulated by chewing. The amount of bone loss varies from person to person. It can be so severe that people are unable to wear removable dentures comfortably. This can make eating and speaking difficult and may affect confidence, well-being, and social relationships. Dental implants can help address these problems.

What are dental implants, and how do they help?

Standard dental implants are short screws (0.5 to 2 cm long) placed into the jawbone which act like artificial tooth roots. Dentists can attach permanent ('fixed') teeth to them, or removable or semi-removable dentures.

Cheekbone implants ('zygomatic implants') are longer: about 3 to 5 cm. They are used when the upper jawbone has shrunk so much that there is not enough bone left to hold standard implants securely. Cheekbone implants almost always support fixed teeth, not removable dentures.

With cheekbone implants, fixed teeth can often be placed on the same day as the implants. This is because they avoid the need for jawbone grafting, which is when a surgeon adds new bone (or bone-like material) to the jaw so it can support an implant. Bone grafting is complex, expensive, and time-consuming. It may require several surgeries and take more than a year before replacement teeth can be fitted.

Cheekbone implant screws can be placed using rotary drills or piezoelectric surgery, which uses sound vibrations to cut bone and may be less traumatic. Cheekbone dental implants can lead to complications, and the potential benefits need to be weighed against these risks.

What did we want to find out?

We wanted to find out if, compared to standard implants in grafted jawbone, cheekbone implants led to fewer:

• replacement teeth failures – where the teeth break or loosen;

• implant failures – where the implant screws become unstable;

• complications.

We also wanted to find out:

• which type of implant led to higher satisfaction and faster rehabilitation;

• which method for implant site preparation was faster for surgeons and preferred by people undergoing treatment.

What did we do?

We searched for studies that compared:

• cheekbone implants to standard implants in grafted jawbone in people with severe bone loss in the upper jaw; or

• different ways to prepare the cheekbone implant site.

We compared and summarised the study results, and rated our confidence in the evidence, based on factors such as study methods.

What did we find?

We found two small studies that followed participants for three years. They were both carried out in Europe. One compared cheekbone implants to standard implants in grafted bone, and involved 71 people. The other compared piezoelectric surgery to rotary drills for preparing the implant site, and involved 20 people.

Main results

After three years, compared to standard implants in grafted jawbone, cheekbone implants:

• may have led to fewer replacement teeth failures: for every 100 people treated, about 12 fewer people with cheekbone implants had problems with their replacement teeth;

• probably had fewer implant failures: for every 100 people treated, about 22 fewer people with cheekbone implants had problems;

• probably led to more complications: roughly four-fifths of people with cheekbone implants had complications, compared to about half of the people with standard implants.

We do not know which procedure left people more satisfied. However, people who received cheekbone implants probably received their new teeth much sooner – in a little more than 1 day – than those with standard implants, where it typically took 444 days before replacement teeth could be fitted.

In the study that compared two methods (piezoelectric surgery versus rotary drills) for preparing the implant sites, after three years, there may be little or no difference between them in terms of:

• replacement teeth failures;

• implant failures;

• complications;

• people's preference.

However, rotary drills probably made implant placement faster by more than 9 minutes.

What are the limitations of the evidence?

Our confidence in the evidence ranged from very low to moderate. We found only two studies testing these approaches, and they involved few people. More well-designed studies are needed to identify the best methods.

How up to date is this evidence?

The evidence is current to November 2025.

Key messages

• Giving adults smaller amounts of red blood cells during a blood transfusion may not make any difference to their risk of dying, how long they stay in hospital, or any unwanted treatment effects, compared with giving larger amounts of red blood cells, but the evidence from some studies is unclear.

• At the same time, it may reduce the amount of blood given for one transfusion, which saves money and means that more blood is available for other people who need a transfusion.

• We did not find enough information to know whether it is better to give children or babies who need a transfusion larger or smaller amounts of red blood cells.

What is a blood transfusion?

Many people in hospital do not have enough healthy red blood cells to carry oxygen around their body, which can make them tired, weak, and short of breath. This is called anemia, and it can be treated in several different ways. Sometimes, people may need to have extra blood added to their own (a blood transfusion). Red blood cells are an important part of the extra blood. However, there is not much information available for doctors about the amount of red blood cells (the ‘red blood cell volume’) that should be given.

The volume of red blood cells in the blood can be measured in units (e.g. one unit of red blood cells) for adults. For babies and children, doctors give a certain amount depending on how much the child weighs. For example, they might give 20 mL of red blood cells for every kg the child weighs (20 mL/kg). Because blood transfusion involves some risks and there is a limited amount of blood available, doctors have been recommended to give the smallest effective amount to adults. Currently, it is recommended that one unit of red blood cells (rather than two units) is given to adults in hospital who are not bleeding and whose condition is not changing quickly. In children and newborns, there is no recommended standard amount for transfusion.

What did we want to find out?

We wanted to find out if it is safe to give a smaller amount of red blood cells to adults, children and newborns in hospital who need a blood transfusion. We looked at whether people stayed in hospital longer, had any unwanted effects from the treatment, any problems with their heart, breathing or other body systems, if they started bleeding again, and how many red blood cells were given.

What did we do?

We searched for studies that compared two different volumes of red blood cell transfusion for people in hospital. For example, two units of red blood cells compared with one unit of red blood cell for adults, or 20 mL/kg compared with 10 mL/kg for children. We compared and summarized the results of the studies and rated our confidence in the evidence, based on factors such as study methods and sizes.

What did we find?

We found 12 studies that involved 5478 people. Nine studies included a total of 1945 adults; one included women needing a blood transfusion after childbirth and the others included people who needed a blood transfusion because of blood cancer. One study included 3199 anemic children aged two months to 12 years, and two studies included a total of 334 newborn babies born before 32 weeks of pregnancy.

In adults, when comparing one unit of red blood cells to two units of red blood cells for a single blood transfusion, there may be no increased risk of dying, no increase in length of hospital stay, and no increase in any unwanted effects of the transfusion, blood clots (thrombosis) or starting to bleed again. However, the evidence is unclear.

What are the limitations of the evidence?

We have little confidence in most of the evidence we found because there were not enough studies to be certain about the results. In some of the studies, it is possible that people were aware of which treatment patients were getting. People in some of the studies were not randomly placed into the different treatment groups. This means that differences between the groups could be due to differences between people rather than between the treatments.

How up to date is this evidence?

The evidence is current to August 2025.

Key messages

• Non-medicinal measures (e.g. visiting restrictions or regular testing) may prevent SARS-CoV-2 infections (causing COVID-19 disease) in residents and staff in long term care facilities, but we have concerns about the reliability of the findings.

• More high-quality studies on real-world experiences are needed, in particular.

• More research is also needed on measures in facilities where most residents and staff are vaccinated, as well as regions other than North America and Europe.

What are non-medicinal measures?

Non-medicinal measures are ways of preventing or reducing disease without using medicine, such as vaccines. These include controlling people's movements and contacts, using personal protective equipment (PPE), or regular testing for infection.

SARS-CoV-2 is very infectious. Elderly or disabled people, who live in care homes (long-term care facilities), are vulnerable to infection because they live in close contact with other people, with carers and visitors entering and leaving the facility. Due to age and underlying health conditions, care home residents have an increased risk of becoming seriously ill with COVID-19 and dying from the disease.

What did we want to find out?

We wanted to find out how effective non-medicinal measures are in preventing residents and staff in long-term care facilities from becoming infected with SARS-CoV-2 and in reducing the spread of the infection. We focused on all types of long-term care facilities for adults, such as nursing homes for the elderly and skilled nursing facilities for people living with disabilities.

What did we do?

We searched for studies that investigated the effects of non-medicinal measures in long-term care facilities. To be included, studies had to report how many infections, hospitalisations or deaths the measures prevented in residents or staff, or whether the measures prevented the introduction of the virus into the facilities or prevented outbreaks within facilities. We included any type of study, including observational studies that used ‘real-world’ data, or modelling studies based on assumed data from computer-generated simulations.

What did we find?

We found 22 studies, 11 observational and 11 modelling studies. All studies were conducted in North America or Europe.

There were four main types of measures.

1.Entry regulation measures to prevent residents, staff or visitors introducing the virus into the facility. Measures included staff confining themselves with residents, quarantine for newly-admitted residents, testing new admissions, not allowing the admission of new residents, and preventing visitors from entering facilities.

2. Contact-regulating and transmission-reducing measures to prevent people passing on the virus. Measures included wearing masks or PPE, social distancing, extra cleaning, reducing contact between residents and among staff, and placing residents and staff in care groups and limiting contact between groups.

3. Surveillance measures designed to identify an outbreak early. Measures included regular testing of residents or staff regardless of symptoms, and symptom-based testing.

4. Outbreak control measures to reduce the consequences of an outbreak. Measures included isolation of infected residents, and separating infected and non-infected residents or staff caring for them.

Some studies used a combination of these measures.

Main results

Entry regulation measures (4 observational studies; 4 modelling studies)

Most studies showed that such measures were beneficial, but some studies found no effects or unwanted effects, such as depression and delirium among residents in the context of visiting restrictions.

Contact-regulating and transmission-reducing measures (6 observational studies; 2 modelling studies)

Some measures may be beneficial, but often the evidence is very uncertain.

Surveillance measures (2 observational studies; 6 modelling studies)

Routine testing of residents and staff may reduce the number of infections, hospitalisations and deaths among residents, although the evidence on the number of deaths among staff was less clear. Testing more often, getting test results faster, and using more accurate tests were predicted to have more beneficial effects.

Outbreak control measures (4 observational studies; 3 modelling studies)

These measures may reduce the number of infections and the risk of outbreaks in facilities, but often the evidence is very uncertain.

Combination measures (2 observational studies; 1 modelling study)

A combination of different measures may be effective in reducing the number of infections and deaths.

What are the limitations of the evidence?

Our confidence in these results is limited. Many studies used mathematical prediction rather than real-world data, and we cannot be confident that the model assumptions are accurate. Most observational studies did not use the most reliable methods. This means we cannot be confident that the measure caused the effect, for example, that testing of residents reduced the number of deaths.

How up to date is this evidence?

This review includes studies published up to 22 January 2021.

Key messages

• Current care for preterm infants (babies born early) who are at risk of respiratory distress syndrome (lung problems causing breathing difficulties) usually involves an intervention called continuous positive airway pressure (CPAP). In these infants, giving surfactant (a coating for the lungs to make breathing easier) as a preventive measure may make little or no difference to the risk of chronic lung disease (needing long-term oxygen support), and probably slightly increases the risk of death, compared with giving surfactant only to those babies who actually develop breathing problems.

• In contrast, for preterm infants not receiving CPAP, giving surfactant preventively reduces the risk of death and may slightly reduce the risk of pneumothorax (lung injury), compared to giving surfactant only if and when breathing problems develop.

• In our view, giving all preterm infants CPAP at birth and then giving surfactant only to those who have symptoms of breathing problems may be the best approach to managing the risk of respiratory distress syndrome.

What is a preterm infant?

A preterm infant is a baby who is born early, having spent less than 37 weeks in the womb.

What is surfactant?

Surfactant is a mixture of fats and proteins produced by the body that coats the inside of the alveoli (tiny air sacs in the lungs) and helps the lungs work properly. Some preterm infants are born before their lungs produce enough surfactant, which makes it difficult for their lungs to stay open and get enough oxygen. This is called respiratory distress syndrome (RDS). Surfactant replacement therapy can be used to replace the missing surfactant.

What did we want to find out?

We wanted to know whether to give surfactant to preterm babies, preventively, that is, immediately after birth before any breathing problems occur ('prophylactic surfactant'), or only if the baby has symptoms ('selective surfactant').

Studies conducted 30 years ago suggested that preterm infants who were intubated (breathing tube placed in the airway, down the throat, and into the lungs) and were treated with prophylactic surfactant (before breathing problems occurred) had fewer RDS complications such as chronic lung disease or pneumothorax.

Newer studies, however, suggest that a better treatment approach may be using breathing support (i.e. continuous positive airway pressure (CPAP)) through a mask or nose prongs placed on the baby’s face, and applying surfactant only if breathing problems develop. This avoids putting tubes into the windpipe and giving medicines that may be unnecessary.

What did we do?

We searched for studies that tested the prophylactic versus selective use of surfactant. We combined results for infant mortality (death) within the first 28 days of life (the neonatal period) and the development of more severe lung disease that may require long-term oxygen supplementation (chronic lung disease). We looked at whether results varied if preterm babies were given CPAP to support their breathing at birth or if their mothers had received medication, such as steroids, which help the baby’s lungs mature during the prenatal period.

What did we find?

We identified 10 relevant studies that involved 3151 babies.

When considering all studies together, prophylactic surfactant probably results in little to no difference in the risk of chronic lung disease (CLD) compared to selective use of surfactant (5 studies, 1874 babies) and may slightly reduce the risk of pneumothorax (8 studies, 3094 babies). It may make little to no difference to the risk of moderate-to-severe neurodevelopmental impairment (problems with brain development) (1 study, 976 babies).

Studies where infants received CPAP showed that prophylactic use of surfactant probably slightly increases the risk of mortality compared to its selective use as a treatment once RDS symptoms have developed, and it may make little to no difference to the risk of CLD. The evidence for pneumothorax is very uncertain.

In contrast, studies where infants did not receive CPAP showed that prophylactic use of surfactant reduces mortality risk compared with giving surfactant selectively after RDS develops, and it may reduce the risk of pneumothorax slightly. The evidence for CLD is very uncertain.

In our view, the review findings favor early stabilization of breathing with CPAP for all preterm infants at risk of developing RDS and then selective use of surfactant for preterm infants showing RDS symptoms.

What are the limitations of the evidence?

People running the studies and families of babies taking part may have known which treatment was being given, which may have influenced measurements.

Not all studies provided information about everything of interest. For example, some studies did not measure longer-term infant development.

The studies did not provide enough reliable information for us to understand the effects of the age of the babies (born before 28 weeks or closer to their due date), mothers receiving steroids before giving birth, or different methods of giving the surfactant.

How up to date is this evidence?

This review updates a previous 2012 Cochrane review. The evidence is current to 31 January 2025.

Key messages

• In people with hip osteoarthritis (a condition where the cartilage cushioning the hip joint wears down, causing pain and stiffness), exercise compared with dummy treatment may have little to no effect on pain, but may improve joint function slightly. Compared with no treatment, usual care, or limited education, exercise probably leads to small improvements in pain and joint function, although these may not be noticeable in daily life. Adding exercise to another treatment probably has little to no effect on pain and joint function.

• We are unsure if exercise leads to more unwanted events than dummy treatment or no treatment/usual care in people with hip osteoarthritis. When added to another treatment, exercise probably reduces the risk of unwanted events slightly.

• We need more and better-designed studies that focus specifically on people with hip osteoarthritis to draw more solid conclusions about the effects of exercise.

What is hip osteoarthritis?

Hip osteoarthritis occurs when the cartilage cushioning the hip joint wears down. It is a common, long-lasting condition. People with hip osteoarthritis often experience joint pain and stiffness. They may have difficulty performing everyday activities such as climbing stairs and getting up from a chair.

How might exercise help?

Exercise aims to improve muscle strength, joint movement, and self-belief in the capacity to perform certain activities. These effects should reduce pain and disability, helping people to better manage their symptoms. Exercise programmes vary considerably, with a wide range of different exercise types, intensities, and delivery methods.

What did we want to find out?

We wanted to know whether exercise reduces pain and improves joint function and well-being for people with hip osteoarthritis. We also wanted to know if people in the studies who exercised were more likely to report treatment success and less likely to drop out of the study or have unwanted events.

What did we do?

We searched for studies that assessed the effects of exercise in adults with hip osteoarthritis. Exercise programmes had to be land-based, planned, structured, and repetitive. We compared and summarised the results of the studies and rated our confidence in the evidence, based on factors such as study methods and how consistent the results were across the studies.

What did we find?

We found 18 studies with 1368 adults. Participants were mostly women (63%) and aged between 53 years and 74 years. Studies took place in Australia, Canada, Denmark, Finland, Germany, Ireland, Italy, the Netherlands, New Zealand, Norway, Sweden, and the USA. Exercise programmes lasted between two and 52 weeks. Most studies measured results immediately after the exercise programme; some studies also reported longer-term results.

Exercise compared with dummy treatment (2 studies, 123 adults)

• Exercise may have little to no effect on pain.

• Exercise may improve joint function slightly. On a scale of 0 to 100, where a 13-point difference is a meaningful change, people who exercised scored 7.4 points better than those on dummy treatment.

• No studies reported well-being or treatment success.

• People who exercise may be no more likely to stop treatment.

• We are unsure if there are more unwanted events with exercise.

Exercise compared with no treatment, usual care, or limited education (10 studies, 494 adults)

• Exercise probably improves pain and joint function, but only slightly. On a scale of 0 to 100, people who exercised scored 7.2 points better for pain (where a 12-point difference is a meaningful change) and 8.8 points better for joint function (where a 13-point difference is a meaningful change), compared with people receiving no treatment, usual care, or education.

• Exercise probably has little to no effect on well-being.

• People who exercise may be no more likely to report treatment success or to stop treatment.

• We are unsure if there are more unwanted events with exercise.

Exercise plus another treatment compared with the other treatment alone (7 studies, 751 adults)

• Exercise probably has little to no effect on pain, joint function, well-being, or whether people stop treatment.

• Exercise may have little to no effect on whether people report treatment success.

• People who exercise probably have slightly fewer unwanted events.

What are the limitations of the evidence?

Because the studies included people with different levels of hip pain and disability, the results are likely to apply to most adults with hip osteoarthritis. However, there were very few younger adults in the studies, so the findings may not apply as well to younger people.

In most studies, people knew whether they were receiving exercise, and pain and physical function were self-reported. This may make exercise appear more effective than it really is.

We are moderately confident in some of the results and less confident in others. This is because many studies were small, and the results varied across the studies. Further research may change the results.

How up to date is this evidence?

This review updates our previous review published in 2014. The evidence is up to date to February 2025.

Key messages

• When types of minimally invasive glaucoma surgery that act on Schlemm's canal (called SC-MIGS) are compared with non-minimally invasive (conventional) glaucoma surgeries, there may be little to no difference in the proportion of people who are medication-free in the short, medium, or long term.

• More people may be medication-free in the short term after SC-MIGS compared with cataract surgery alone; however, there may be a higher risk of bleeding.

• SC-MIGS may reduce eye pressure more than laser treatment.

What is open-angle glaucoma?

Open-angle glaucoma is a common eye condition in which there is damage to the optic nerve, which usually progresses slowly over many years. 'Open-angle' means that the eye's natural drainage channels are not blocked but do not drain fluid effectively. This can increase pressure in the eye and damage the optic nerve, causing vision loss, often affecting the peripheral vision before the central vision. Some people develop glaucoma despite normal eye pressure and require further pressure-lowering treatment.

How is open-angle glaucoma treated?

Treatment includes eye drops, laser treatment, and surgery. Surgery is usually used if eye drops or laser treatment has not worked.

Standard surgery usually involves bypassing the drainage channels of the eye to improve fluid drainage. It may be performed under local anaesthetic (while the patient is awake) or general anaesthetic (while the patient is asleep). Complications, such as bleeding in the eye or very low eye pressure, can occur, and some people may need repeat surgery.

One alternative to standard surgical procedures is minimally invasive glaucoma surgery. Some of these procedures act on a part of the drainage network of the eye called Schlemm's canal.

What did we want to find out?

We wanted to find out whether these types of minimally invasive glaucoma surgery that specifically act on Schlemm's canal (called SC-MIGS below) help improve outcomes and whether they cause any harmful or unwanted effects in people with open-angle glaucoma. We were particularly interested in:

• the likelihood of people not needing eye drops;

• the likelihood of progressive visual field loss (loss of central or side areas of vision);

• changes in eye pressure;

• changes in the number of eye drops used following surgery;

• any unwanted effects or side effects.

What did we do?

We searched for studies that compared SC-MIGS to other types or combinations of surgery, or to laser treatment. We compared and summarised the results of the studies and rated our confidence in the evidence, based on factors such as study methods and sizes.

What did we find?

We found 10 studies that involved 639 people (639 eyes). People in the studies were men and women aged between 41 and 74 years. The studies took place in Italy, Japan, India, the USA and Latin America, Poland, Canada, and China. The studies compared:

• SC-MIGS with conventional glaucoma surgery;

• SC-MIGS plus cataract surgery with cataract surgery alone;

• SC-MIGS with types of minimally invasive glaucoma surgery that don’t act on Schlemm’s canal;

• one type of SC-MIGS with another type of SC-MIGS;

• SC-MIGS with laser procedures.

SC-MIGs versus standard glaucoma surgery (3 studies, 99 people)

There may be no difference between SC-MIGS and standard glaucoma surgery in the likelihood of people not needing eye drops to control their glaucoma, having progressive visual field loss, or in the average change in eye pressure, number of eye drops people used following surgery, or unwanted effects.

SC-MIGS plus cataract surgery versus cataract surgery alone (2 studies, 132 people)
People may be less likely to need eye drops to control their glaucoma, average change in eye pressure may be greater, but there may be more risk of bleeding with SC-MIGS plus cataract surgery compared with cataract surgery alone.

SC-MIGS versus types of minimally invasive glaucoma surgery that don’t act on Schlemm’s canal (2 studies, 238 people)
There may be no difference between these two treatments in how likely people are not to need eye drops to control their glaucoma, number of eye drops used following surgery, or need for further surgery.

One type of SC-MIGS versus another type of SC-MIGS (2 studies, 140 people)
There is likely no difference between one type of SC-MIGS and another in any of the things we wanted to find out.

SC-MIGS versus laser procedures (1 study, 30 people)
Eye pressure may be reduced more, but there may be more risk of bleeding with SC-MIGS compared with laser procedures.

What are the limitations of the evidence?

Our confidence in the evidence is very limited because there were not many studies, and they did not include many people. The studies did not use the best methods, so their results may not be reliable. Also, the studies each focused on a small group of people of the same ethnicity or nationality, so we don’t know how their findings will apply to different groups of people.

How up to date is this evidence?

The evidence is up to date as of October 2025.

Key messages

• Transfusing platelets (giving blood cells that form clots to stop bleeding) at higher platelet count levels (when there is less risk of bleeding) does not seem to decrease death or bleeding.

• Transfusing at a lower platelet level likely reduces platelet transfusions without increasing the risk of death or bleeding.

• Future studies are needed to know the exact platelet count at which babies should be given platelets based on their individual risks for bleeding.

What are platelets?

Platelets are blood cells that form clots to stop bleeding. A low platelet level (platelet level on a blood test of less than 150,000 platelets per microliter) is known as thrombocytopenia. Thrombocytopenia is common in premature or sick newborn babies and can increase their risk of bleeding.

How is thrombocytopenia treated?

Thrombocytopenia is commonly treated by giving platelets from donated blood. Platelets are given directly into a vein in a procedure called a transfusion. However, platelet transfusion itself may increase the risk of death and other harms, and experts in the field differ about whether a platelet transfusion should be done when babies' platelets are lower versus when they are higher.

What did we want to find out?

We wanted to find out if it is better to give platelets when platelet levels are higher compared to when they are lower in babies with thrombocytopenia. We wanted to know the effects of transfusion on death and bleeding, and on babies' development.

What did we do?

We searched for studies that looked at giving platelet transfusions to babies with thrombocytopenia. We compared and summarized the results of the studies and rated our confidence in the evidence based on factors such as study methods and sizes.

What did we find?

We found three studies (856 babies) that compared different platelet levels for treatment with a platelet transfusion. Two studies took place in high-income countries (the USA, Canada, the UK, Ireland, and the Netherlands), and one in a lower-middle-income country (India). One study looked at whether it was better to give a platelet transfusion when the platelets in babies were just a little low (mild thrombocytopenia) compared to if the platelets were low (moderate thrombocytopenia) (mild thrombocytopenia range). Another study looked at whether it was better to give a platelet transfusion when the platelets were just low (moderate thrombocytopenia) compared to if the platelets were very low (severe thrombocytopenia) (moderate thrombocytopenia range). The third study looked at whether it was better to give a platelet transfusion when the platelets were very low (severe thrombocytopenia) compared to if the platelets were extremely low (very severe thrombocytopenia) (severe thrombocytopenia range).

We found no difference between lower and higher platelet levels in death:

• mild thrombocytopenia range (1 study, 152 babies);

• moderate thrombocytopenia range (1 study, 44 babies);

• severe thrombocytopenia range (1 study, 656 babies).

We found no difference between lower and higher platelet levels in bleeding:

• mild thrombocytopenia range (not reported);

• moderate thrombocytopenia range (1 study, 44 babies);

• severe thrombocytopenia range (1 study, 658 babies).

We found that transfusing at a lower platelet level likely reduces platelet transfusions without increasing the risk of death or bleeding:

• mild thrombocytopenia range (1 study, 152 babies);

• moderate thrombocytopenia range (not reported);

• severe thrombocytopenia range (1 study, 659 babies).

What are the limitations of the evidence?

We have only moderate to very low confidence in the evidence. Our confidence is limited because the people doing the studies knew which treatment they were giving, and we only found three studies, all of which differed in the platelet levels studied. The studies also differed in whether they included babies who were already bleeding. The largest study did not include babies with bleeding.

How up-to-date is this evidence?

The evidence is current to June 2025.

Key messages

• In children and adolescents with neurological impairment, we are very uncertain about the effect of adding an antireflux procedure (fundoplication) on symptoms of gastro-esophageal reflux, death, and major surgical complications (perforation or peritonitis), when compared to gastrostomy alone or gastro-jejunal tubes.

• The addition of an antireflux procedure when performing a gastrostomy may result in little to no difference in the number of hospitalizations for respiratory disease (pneumonia).

• Further research is necessary. Better designed observational studies could help to improve our confidence in the evidence.

Why treat gastro-esophageal reflux in children with neurological impairment?

Because of damage to their central nervous system, children with neurological impairment are prone to vomiting and feeding difficulties, and are at increased risk of dying from lower respiratory tract infections, such as pneumonia. The treatment of gastro-esophageal reflux (where the stomach acid frequently flows back into the esophagus [food pipe] and causes heartburn) aims to prevent these symptoms and complications.

How is gastro-esophageal reflux that doesn't get better with medicine treated in these children?

The two most common procedures used to treat or prevent the consequences of gastro-esophageal reflux (antireflux procedures) in children with neurological impairment are fundoplication, where the stomach is wrapped around the lower esophageal sphincter, and the placement of a gastro-jejunal tube, which passes a feeding tube from the stomach into the small intestine (jejunum). These procedures differ in their surgical risks: fundoplication is more complex, requires more anesthesia (sedation), and involves a longer recovery period. Gastro-jejunal tubes are easier to place but require continuous feeding, which may increase caregiving needs and reduce the quality of life of children and their families.

What did we want to find out?

We wanted to find out if a fundoplication or passing of a gastro-jejunal tube is better than a gastrostomy (feeding tube entered directly into the stomach) alone to improve symptoms of gastro-esophageal reflux and number of deaths.

We also wanted to know if fundoplication caused more major surgical complications, and if antireflux procedures could prevent respiratory morbidity (pneumonia) in the long term.

What did we do?

We searched for studies comparing the two antireflux procedures (fundoplication or gastro-jejunal tubes) versus gastrostomy alone or versus each other in children and adolescents with neurological impairment.

We compared and summarized the results of the studies and rated our confidence in the evidence based on factors such as study methods and sizes.

What did we find?

We found 11 studies involving a total of 3122 children with neurological impairment. The children were followed up for between 9 months and 12 years.

Main results

Gastrostomy plus fundoplication compared to gastrostomy alone for gastro-esophageal reflux

• We are very uncertain about the effect on symptoms of gastro-esophageal reflux (3 studies, 180 children), death (3 studies, 415 children), major surgical complications within the first 30 days after surgery (3 studies, 412 children), and emergency department visits after surgery (1 study, 130 children).

• Gastrostomy plus fundoplication may result in little to no difference in number of hospitalizations for respiratory disease (pneumonia) (1 study, 2054 children).

• No studies looked at child's quality of life for this comparison. Length of stay after surgery was reported but not compared for gastrostomy plus fundoplication versus gastrostomy alone.

Gastro-jejunal tubes compared to gastrostomy alone for gastro-esophageal reflux

• We are very uncertain about the effect on death (1 study, 50 children) and child's quality of life (1 study, 50 children).

• No studies looked at symptoms of gastro-esophageal reflux, major surgical complications, length of stay after surgery, number of hospitalizations for respiratory disease (pneumonia), or number of emergency department visits after surgery for this comparison.

Gastrostomy plus fundoplication compared to gastro-jejunal tubes for gastro-esophageal reflux

• We are very uncertain about the effect on symptoms of gastro-esophageal reflux (3 studies, 1531 children), death (3 studies, 418 children), and major surgical complications during the first 30 days after surgery (2 studies, 190 children).

• Gastrostomy plus fundoplication may result in little to no difference in number of hospitalizations for respiratory disease (pneumonia) (1 study, 228 children).

• No studies looked at number of emergency department visits after surgery or child's quality of life for this comparison. Length of stay after surgery was reported but not compared for gastrostomy plus fundoplication versus gastro-jejunal tubes.

What are the limitations of the evidence?

We have little to very low confidence in the evidence. There were many problems with the methods of the included studies, such as incomplete reporting, differences among the participants in each group, and small study sizes that prevented the recording of events that don't occur very often during the follow-up period. Also, the included studies did not report all the outcomes we were interested in.

How up-to-date is this evidence?

The evidence is current to September 2024.

Key words: neurological impairment, children, adolescents, gastrostomy, tubes, fundoplication, gastro-jejunal, gastro-esophageal reflux disease, gastro-oesophageal reflux, GERD.

Key messages

• A common treatment for inflammatory arthritis is a medicine that alters the immune system and so reduces inflammation (DMARDs). Stopping DMARD treatment around the time of surgery to treat an affected joint may increase people's risk of experiencing a flare (worsening symptoms), but may make no difference to their likelihood of developing an infection or experiencing an unwanted event after the operation.

• More studies are needed to assess the effects of stopping or continuing DMARDs on disease activity, function, and quality of life.

What is inflammatory arthritis?

Inflammatory arthritis (including rheumatoid arthritis, psoriatic arthritis and axial spondyloarthritis) is a chronic (persistent) condition that causes abnormal inflammation in the joints, ligaments (connective tissue), and tendons (which attach muscle to bone), resulting in pain, stiffness, swelling, and loss of function. It is a progressive disease that, if uncontrolled, may result in irreversible joint damage, significantly impaired function and quality of life, and inability to work.

How is inflammatory arthritis treated?

Inflammatory arthritis is treated with medicines called 'disease-modifying anti-rheumatic drugs' or 'DMARDs' that control the inflammation, thus reducing symptoms (pain, swelling, or stiffness) and preventing joint damage.

However, since DMARDs alter the immune system, they may also increase the risk of infection, including surgical site infections.

What did we want to find out?

We aimed to find out about the benefits and risks of stopping compared to continuing DMARDs around the time of surgery in people with inflammatory arthritis. We were interested in:

• flare (worsening symptoms or increased disease activity)

• infections after surgery

• infection in the artifical (prosthetic) joint

• disease activity score (from questionnaires)

• function (measured by questionnaire)

• unwanted events (such as wound problems or complications from surgery)

What did we do?

We searched for studies that investigated the effects of stopping or continuing DMARDs around the time of surgery. People in the studies had to be adults with inflammatory arthritis. They could be undergoing any type of surgery except for emergency or unplanned surgery.

We compared and summarised the results and rated our confidence in the evidence, based on factors such as study methods and sizes.

What did we find?

We found three studies with 306 people. All had rheumatoid arthritis and had been taking conventional synthetic (man-made) DMARDs for more than 10 years. They were all undergoing planned orthopaedic surgery, that is, surgery on bones, joints and tendons. Their average age was 56 years, and 83% were women. Studies were conducted in France, the UK and Japan. One trial reported receiving funding from a foundation unrelated to commercial purposes and two studies did not report funding sources.

Main results

Compared with continuation of DMARDs during surgery:

• discontinuation may increase the risk of flare: 37% more people had a flare or worsening of the underlying disease activity when stopping DMARDs during surgery compared to those who continued with DMARDs (2 studies, 224 people). Out of 100 people who continued DMARDs, none reported a flare, compared to 37 out of 100 people who stopped DMARDs;

• discontinuation may have little or no effect on postoperative infections: 0% fewer people had postoperative infections when stopping DMARDs (3 studies, 306 people). Out of 100 people who continued DMARDs, 3 had postoperative infections compared to 3 out of 100 people who stopped with continuation of DMARDs.

• only one study (64 people) reported postoperative joint infections, with no infections in people who continued or stopped DMARDs.

• discontinuation may have little or no effect on total unwanted events, such as problems with the wound (3 studies, 306 people) or serious unwanted events, such as need for further surgery or procedures (2 studies, 242 people); 9% more people reported an unwanted event and 2% more people reported a serious unwanted event when stopping DMARDs.

What are the limitations of the evidence?

We have little confidence in the evidence because we found only three studies, and people in the studies were aware of which treatment they were getting. This may affect their reporting of disease flare, unwanted events and serious unwanted events, but it probably does not affect how they reported postoperative infections. None of the studies measured prosthetic joint infections, disease activity and function, so we don't know the effect of stopping DMARDs around the time of surgery on them. We only found studies on one type of DMARD.

How up to date is this evidence?

The evidence is up-to-date to 24 March 2025.

Editorial note: This is a living systematic review. We search for new evidence approximately yearly and update the review when we identify relevant new evidence. Please refer to the Cochrane Database of Systematic Reviews for the current status of this review.

Key messages

• Cyclosporine A (CsA) eye drops, including Restasis, Cequa, and other newer versions, may improve some signs and symptoms of dry eye compared with their "vehicle" (inactive eye drops containing no CsA) or artificial tears. However, we have only moderate or low confidence in these results.

• People who used CsA eye drops were slightly more likely to stop treatment due to unwanted effects, but overall, few people withdrew from studies for this reason.

• We need higher-quality, longer-lasting studies that look at the benefits and harms of CsA eye drops in people with dry eye disease of varying seriousness.

What is dry eye disease?

The eye's surface is moist and covered with a thin film of tears. When problems arise with the tear film, people can develop a common condition known as dry eye disease. People with dry eye may have eye discomfort (such as a feeling of grittiness), burning sensations, light sensitivity, and blurred or fluctuating vision, which can affect driving and reading and lower their quality of life.

Some people develop dry eye disease because of another health condition or as a result of eye treatment, while in others, the cause is unknown. Many cases of dry eye are linked to health conditions that cause inflammation (swelling and irritation) in and around the eyes.

What are cyclosporine A (CsA) eye drops?

CsA eye drops are anti-inflammatory medicines that aim to reduce eye surface inflammation, improve tear production, and prevent eye surface damage. CsA eye drops come in different strengths, typically ranging from 0.05% (lower strength) to 2% (higher strength).

What did we want to find out?

We wanted to find out whether CsA eye drops work better than eye drops containing all the same ingredients except cyclosporine ('vehicle eye drops' or just 'vehicle'), artificial tears, or another type or strength of CsA to improve:

• dry eye symptoms;

• eye surface health;

• tear production;

• how well tears stay on the eye (tear film stability).

We also wanted to find out whether people using CsA eye drops were more likely to stop using them because of unwanted effects than those using vehicle, artificial tears, or CsA drops of a different type or strength.

What did we do?

We searched for studies comparing CsA eye drops with vehicle, artificial tears, or another CsA formulation or strength. We compared and summarized the results of these studies and rated our confidence in the evidence based on factors such as study methods and the number of people in the studies.

What did we find?

We found 58 studies involving 10,225 people with dry eye disease. China, South Korea, and the USA contributed 35 of the 58 studies (60%). Participants were typically around 53 years old. Women made up about 80% of participants overall, though proportions varied across studies from 46% to 100%. Pharmaceutical companies funded or sponsored nearly half of the studies. Non-profit organizations supported 10 studies, six studies had no external funding, and 14 studies did not report funding sources.

We grouped the studies into six groups based on CsA strength and the comparison treatments. Here, we focus on three groups with the most data.

Main results

Comparison 1

Compared to vehicle, artificial tears, or both, using CsA 0.05% eye drops for three months may have slightly improved people's:

• dry eye symptoms (13 studies, 1396 people);

• eye surface health (7 studies, 1355 people).

Results for tear production and tear film stability were inconsistent, preventing us from drawing firm conclusions.

Comparison 2

People using CsA eye drops of different strengths and types may have had slightly better tear production than those using CsA 0.05% eye drops after three months, but the difference between groups was small (5 studies, 572 people). We are uncertain about the effects on:

• dry eye symptoms;

• eye surface health;

• tear film stability.

Comparison 3

Compared to vehicle, artificial tears, or both, using CsA 0.1% eye drops for three months:

• probably improved eye surface health slightly (7 studies, 2940 people);

• may have improved tear production slightly (3 studies, 735 people).

Results for dry eye symptoms were inconsistent, preventing us from drawing firm conclusions. We are uncertain about the effect of CsA 0.1% on tear film stability.

Stopping treatment due to unwanted effects

Compared with people using vehicle eye drops or artificial tears, those using CsA 0.05% or 0.1% were slightly more likely to stop treatment due to unwanted effects. Overall, only a few people stopped treatment because of unwanted effects.

What are the limitations of the evidence?

We have little confidence in the evidence for dry eye symptoms, eye surface health, tear production, and tear film stability because study results were often inconsistent or too uncertain to determine the true size of the effect. Many studies gave unclear or few details about their methods. However, we are moderately confident in the evidence about stopping treatment.

How up to date is this evidence?

The evidence is current to March 2026.

Key messages

• Venepuncture (inserting a needle into a vein) may be less painful than heel lance (using a special device to puncture the skin) for newborn babies when having a blood sample taken, with or without sugar water for pain relief, and may reduce the need for more than one skin puncture when no sugar water is given for pain relief.

• We don’t know which procedure, venepuncture or heel lance, is more painful for newborn babies when sugar is used for pain relief.

• We need better designed and larger studies that fully report their results to show which method is better for babies.

Why do we take blood samples from newborn babies?

In most countries, a blood sample from newborn babies is needed for routine screening tests. Blood samples may also be needed for other blood tests to diagnose and manage diseases, such as jaundice or low blood sugar (hypoglycemia). However, taking blood from babies is painful. Untreated pain in newborn babies may affect their physical growth and brain development. A sweet, sugar solution (such as sucrose) can be given to babies before the procedure, which helps to reduce the pain. It is difficult to measure pain in babies, but we rely on things like facial expressions, crying, and changes in heart rate to assess their pain.

How are blood samples taken?Venepuncture

Venepuncture is a method of taking a blood sample where a person uses a needle to directly puncture the vein. The person who performs this must be well-trained to help ensure they successfully take a blood sample. This technique is often used when 1 mL or more of blood is required, and to prevent repeated punctures to the skin and associated complications (such as skin infections) from heel lance.

Heel lance

Heel lance is the standard way of taking a blood sample from newborn babies in hospitals worldwide. It is commonly used when the blood sample required is less than 1 mL. It does not require special training or expertise for successful use. The procedure involves making a small puncture (between 1 mm and 2 mm deep) on the heel of the foot using a lancet, which is usually an automatic device. Following the puncture, the infant's heel is squeezed to collect the required blood sample, which causes further pain.

What did we want to find out?

We wanted to find out:

• whether venepuncture is less painful than heel lance for newborn babies; and

• which procedure works best first time, so a second skin puncture is not needed.

What did we do?

We searched for studies that compared venepuncture and heel lance for taking blood samples in newborn babies. The studies needed to measure pain responses and the number of skin punctures needed for a successful sample. They also needed to record whether babies were given sugar solution as pain relief. Babies in the studies had to be healthy and born at full term. We compared and summarized the results of the studies and rated our confidence in the evidence, based on factors such as study methods and sizes.

What did we find?

We included eight studies involving 826 full-term newborn babies. Three studies took place in Sweden, and one each in Japan, Malaysia, Nepal, Spain, and the UK.

When no sweet-tasting solution was given:

• venepuncture probably causes less pain compared to heel lance (6 studies, 488 babies); and

• venepuncture probably reduces the need for more than one skin puncture compared to heel lance (6 studies, 397 babies).

When a sweet-tasting solution was given:

• venepuncture may cause less pain compared to heel lance (3 studies, 170 babies); and

• we are uncertain about the effect of venepuncture compared to heel lance on the need for more than one skin puncture (2 studies, 110 babies).

What are the limitations of the evidence?

Our confidence in the evidence was limited, and ranged from moderate to very low. This is because all the studies had limitations in how they were carried out. These included issues with how babies were assigned to each group, how the studies reported their results, and the number of babies included in the studies. In addition, the people doing the procedures knew which sampling method each baby received, which could have influenced how the studies assessed pain.

How up to date is this evidence?

This review updates our 2011 Cochrane review. The evidence is up to date to July 2025.

Key messages

• Most children with growth-plate pain (apophysitis) get better over time. Whether treatment for apophysitis affects short-term pain, physical function, and participation in sport is uncertain.

• It is uncertain whether any treatment is better or worse than any other. The studies were mostly small and had problems with their design. The studies did not always measure unwanted effects from a treatment, and none measured whether any children withdrew from studies because of unwanted effects.

• The evidence is limited. Most studies included more boys than girls, and included children who were more active than average. The children often knew which treatment they were receiving, which may have influenced the results.

What is apophysitis?

During puberty, children sometimes develop pain in the growth plates in their hips, knees, and feet. This pain is usually caused by irritation of the growth plate from repeated stress over time rather than from a single event. The pain is usually short-term and has limited impact on a child's life. However, sometimes the pain can make children limp, play less sport, or be less physically active. The conditions are called different names based on the location of the growth, but are commonly known as 'apophysitis' or 'lower limb apophyseal injuries'.

How is apophysitis treated?

Many different healthcare professionals treat apophysitis, including general practitioners or physicians, physiotherapists, and podiatrists. Many treatments are used, including exercises, medicines, strapping, braces, or devices put in shoes like foot orthoses or heel lifts. Regardless of treatment, apophysitis usually gets better on its own, but in some cases, the pain can last for a long time.

What did we want to find out?

We wanted to find out which treatments reduce pain effectively and to understand how safe these treatments are. We wanted to know which treatments improve how children function physically and take part in sport. We also wanted to know if one treatment worked better than another.

What did we do?

We searched for studies that tested the different types of treatments for apophysitis in the hip, knee, and foot. We did not search for any treatments that involved surgery.

We combined results from studies that tested the same treatments and used similar methods to measure effectiveness. We rated our level of certainty about this evidence.

What did we find?

We found 10 studies that involved 654 children whose average age ranged from 10.3 to 13.3 years. Seven studies focused on heel pain (calcaneal apophysitis), and three focused on pain in the front of the knee (traction apophysitis of the tibial tubercle).

Main results

• One study (23 children) compared a medicine called dexamethasone to a placebo (fake treatment). We are very uncertain whether dexamethasone reduces pain, improves physical function, or helps children return to sport in the short term. Two studies (74 children) reported unwanted effects of medicines, but the evidence was also very uncertain.

• One study (21 children) compared dexamethasone to usual care. We are very uncertain whether dexamethasone reduces pain, improves physical function, or helps children return to sport in the short term. One study (30 children) reported unwanted effects, but the evidence was also very uncertain.

• One study (22 children) compared Kinesio tape to a placebo. We are very uncertain whether Kinesio tape improves pain or physical function in the short term. The study did not report on participation in sport or unwanted effects.

• One study (124 children) compared foot orthoses to heel lifts. It found that there is probably little to no difference between them in terms of pain or physical function in the short term. The study did not report on participation in sport. One study (101 children) reported that there were no unwanted effects.

• One study (43 children) compared heel cushioning to a heel strap. We are very uncertain whether heel cushioning improves physical function or causes any unwanted effects in the short term. The study did not report on participation in sport.

• None of the studies reported whether any children withdrew from the studies because of unwanted effects.

What are the limitations of this research?

• In most studies, the children and carers knew which treatments they were receiving, which may have influenced how they reported improvements. This may affect how reliable the results are.

• Apophysitis often improves on its own as children continue to grow and the growth-plates close, so it is difficult to know if any improvements are due to treatments or natural recovery over time.

• There were many more boys than girls in the studies, and most participants were very physically active. It is unlikely that the condition impacts boys and girls differently, but we are unsure if our findings can be applied to children who are less physically active.

• Finally, most studies had only a small number of participants. This means we cannot be confident in the results and cannot know whether the findings would apply to a wider group of children with apophysitis.

How up to date is this evidence?

The evidence is current to 4 January 2025.

Key messages

- If damaged teeth are filled with a "bulk-fill" (material placed in one large layer) or a standard resin-based composite (RBC) material (placed in multiple layers), there is probably little or no difference in failure between the filling materials.

- Older evidence found that dental amalgam (silver-coloured fillings) may have fewer material failures than RBC. But this finding was based on older studies. Modern RBCs are better, dentists have more experience using them, and there are likely fewer failures.

Why do we need fillings and what are they?

Sometimes the tooth tissue becomes damaged, and holes (cavities) may form in the teeth. Cavities are caused by tooth decay, which is mainly a preventable disease if people adopt good oral hygiene and limit sugary foods and drinks.

When tooth decay leads to permanent damage, dentists may use a filling in the cavity to restore the tooth's shape and function.

What are the different materials used for fillings?

Traditionally, cavities were filled with amalgam. This silver-coloured filling material (made mainly from tin, silver and copper, and mixed with liquid mercury) is low cost and easy to use. But, it is now known that mercury can be harmful to people's health and the environment. At an international meeting, agreement was reached to reduce (or stop) the use of mercury, including in dental fillings. Alternative mercury-free filling materials that can be placed in a single dental appointment include:

- RBC: standard white or tooth-coloured material, applied in layers or in bulk.

- glass ionomer cement (GIC): sometimes used as a temporary filling, not as strong in some situations, so not suitable for chewing surfaces of teeth.

- resin-modified GIC (RMGIC): a hybrid GIC, stronger than GICs but not as strong as RBC.

- compomers: alternative to RMGIC but closer to composite than GICs.

RBC, RMGIC and compomers use an adhesive system to fix the material to the teeth, and need light-curing to harden the material. These are not needed with GIC.

What did we want to find out?

We wanted to find out:

- the benefits of one filling material compared to another for reducing tooth loss (because the filling has failed), failure (when the filling does not work as intended), length of time to failure, and tooth sensitivity after the procedure.

- whether one filling material is more cost-effective than another.

What did we do?

We searched for systematic reviews that looked at materials used for dental fillings in people's permanent back teeth. These reviews collect all the available evidence from published studies and analyse their results. We summarised the results of reviews alongside ratings of confidence (based on factors such as review or study methods and sizes) in the evidence that they reported.

We also looked for economic studies about the cost-effectiveness of these materials.

What did we find?

We found 14 reviews with 57 studies. The studies were undertaken between 1980 and 2023; one study had follow-up for 10 years but generally follow-up was much shorter. Some reviews did not report results for our critical outcomes, and sometimes the same studies were included in more than one review. We prioritised evidence from six reviews, with 25 unique studies.

We found seven studies about cost-effectiveness.

Main results

We found no reviews with evidence for tooth loss or length of time to failure.

- Although amalgam fillings may lead to fewer failures than RBC fillings (8 studies, 3486 fillings), the studies started in the late 1990s. RBCs, and dentists' experience using them, have improved since then. Failure with RBC is more likely to be 5% today (compared to 15% in the older studies). This means that older evidence comparing amalgam with RBC is less useful in modern settings.

- There is probably no difference between bulk-fill and standard RBC at reducing failures (7 studies, 511 fillings). It is likely that most people will experience no tooth sensitivity after either type of RBC material (5 studies, 510 fillings).

- There may be no difference between standard RBC and GIC at reducing failures (1 study, 60 fillings) or tooth sensitivity (4 studies, 311 fillings). RMGIC may be more likely to reduce failures than GIC (1 study, 50 or 38 fillings). We found reviews comparing GIC with amalgam and compomer, but these did not report failures or postoperative sensitivity.

- Most economic studies gave no overall conclusions about the cost-effectiveness of any of the materials. One economic study found that amalgam fillings were likely to last longer and be more cost-effective than RBC fillings, though this was based on an older study (from the late 1990s).

What are the limitations of the evidence?

Most reviews did not meet the highest standards. We judged only two reviews to be well-conducted. However, reviews mostly reported very similar results, even when they included different studies.

We were less confident in the evidence including GIC and RMGIC because studies were often small with few participants.

How current is this evidence?

The evidence is current to April 2025.

Key messages

• Live births are probably slightly decreased with recombinant follicle-stimulating hormone (rFSH) compared to purified (HMG) or highly purified (HP-HMG) human menopausal gonadotropin, but probably increased with rFSH compared to biosimilars (manufactured gonadotropins). There is probably little to no difference in live births between rFSH and purified follicle-stimulating hormone (FSH-HP) or follitropin delta. Clinical pregnancy is probably decreased with rFSH compared to HMG/HP-HMG but probably increased with rFSH compared to biosimilars. There is probably little to no difference in clinical pregnancy with rFSH compared with FSH-HP and follitropin delta.

• OHSS is probably higher with rFSH compared to HMG/HP-HMG and follitropin delta, but there is probably little to no difference with rFSH compared to FSH-HP, and there may be little or no difference with rFSH compared to biosimilars.

• Future research should focus on frozen embryos and cumulative outcomes across multiple frozen embryo transfer cycles to give a more complete picture of treatment effectiveness. Strategies for the prevention and management of OHSS should be prioritised.

What is in vitro fertilisation?

In vitro fertilisation (IVF) is a type of fertility treatment. Ovarian stimulation helps women’s ovaries produce more eggs, which are collected, fertilised and transferred back to the woman. Various hormones, including gonadotropins, are used to stimulate the ovaries:

• purified human menopausal gonadotropin (HMG) and highly purified (HP-HMG);

• purified follicle-stimulating hormone (FSH-P) or highly purified follicle-stimulating hormone (FSH-HP);

• recombinant follicle-stimulating hormone (rFSH) was developed in a laboratory to obtain higher purity of the hormone;

• follitropin delta was developed to work in a more personalised way; and

• 'biosimilars' are manufactured to offer similar results to rFSH at a potentially lower cost.

Ovarian hyperstimulation syndrome (OHSS) is an unwanted and sometimes serious effect of IVF, in which the ovaries swell and fluid can leak into the abdomen.

What did we want to find out?

We wanted to know which gonadotropin produced more live births, caused less OHSS, and led to more 'clinical' pregnancies, confirmed by ultrasound.

What did we do?

We searched for studies that compared one type of gonadotropin with another for women undergoing IVF. Studies could use fresh or frozen embryos.

What did we find?

We included 59 studies with a total of 18,119 women undergoing fertility treatment in a clinic.

rFSH versus HMG/HP-HMG (17 studies, 5639 women)

• Live births probably decrease with rFSH compared to HMG/HP-HMG. If the chance of live birth with HMG/HP-HMG is 28%, then the chance of live birth with rFSH is between 22% and 27%.

• OHSS is probably higher with rFSH compared to HMG/HP-HMG. If the risk of OHSS with HMG/HP-HMG is 2.6%, then the risk of OHSS with rFSH is between 2.9% and 4.7%.

• Clinical pregnancies probably decrease with rFSH compared to HMG/HP-HMG.

rFSH versus FSH-HP (23 studies, 4612 women)

• There is probably little to no difference in live births between rFSH and FSH-HP. If the chance of a live birth with FSH-HP is 27%, then the chance of a live birth with rFSH is between 24% and 31%.

• There is probably little to no difference in OHSS with rFSH compared to FSH-HP. If the risk of OHSS with FSH-HP is 2.5%, then the risk of OHSS with rFSH is between 1.7% and 3.8%.

• There is probably little to no difference in clinical pregnancy between rFSH and FSH-HP.

rFSH versus follitropin delta (7 studies, 3614 women)

• There is probably little or no difference in live births between rFSH and follitropin delta. If the chance of achieving a live birth with follitropin delta is 29%, the chance of a live birth with rFSH is between 24% and 30%.

• OHSS is probably higher with rFSH. If the risk of OHSS with follitropin delta is 4.4%, then the risk with rFSH is between 5.1% and 9.2%.

• There is probably little or no difference in clinical pregnancy between rFSH and follitropin delta.

rFSH versus biosimilars (8 studies, 2870 women)

• Live birth is probably higher with rFSH compared to biosimilars. If the chance of achieving a live birth with biosimilars is 23%, the chance of a live birth with rFSH is between 25% and 32%.

• There may be little or no difference in OHSS between rFSH and biosimilars. If the chance of OHSS with biosimilars is 9.6%, the chance of OHSS with rFSH is between 5.7% and 10%.

• Clinical pregnancy is probably higher with rFSH compared to biosimilars.

What are the limitations of the evidence?

The results mainly reflect IVF cycles with the transfer of a fresh embryo, so the results may be different for frozen embryos. Many studies were sponsored by pharmaceutical companies that manufacture the hormones in question, which could have affected the results. Furthermore, although we excluded some studies by using an integrity checklist, we have some concerns about the trustworthiness of the data from several of the studies.

How up to date is the evidence?

This updates a review previously published in 2011. The evidence is current to March 2025.

Key messages

• In adults who had treatment to reopen a blocked brain artery, lowering blood pressure very intensively (usually below 160 mmHg) does not improve the chances of living independently three months after the stroke compared with standard blood pressure targets (usually below about 180 mmHg). “Living independently” means being able to walk and manage most everyday activities without help.

• Intensive blood pressure lowering probably increases the risk of death and poor recovery after stroke. Differences in brain bleeding between intensive and standard blood pressure strategies were small.

• Most of the people in the studies were treated in hospitals in upper-middle and high-income countries with specialized stroke services. We do not know whether the results would be the same in countries with fewer resources or in all groups of people who have an ischemic (blood blockage) stroke.

What is an ischemic stroke?

A stroke occurs when the blood supply to part of the brain is suddenly interrupted. In an ischemic stroke, this interruption is caused by a blood clot blocking an artery in the brain. It can lead to various neurological problems, such as weakness, decreased sensitivity, lack of coordination, and difficulty speaking.

How is a stroke treated?

Doctors can sometimes reopen the blocked artery using clot-busting medicines given into a vein, or by performing a procedure to remove the clot with a thin tube passing through the blood vessels (often called mechanical thrombectomy). These are called reperfusion treatments because they restore blood flow to the brain.

After reperfusion, many people have very high blood pressure, which can increase the risk of bleeding in the brain. However, lowering blood pressure too much or too quickly can reduce blood flow to brain tissue that is still at risk and be harmful to the brain.

What did we want to find out?

We wanted to find out whether intensive blood pressure lowering after reperfusion treatment, compared with standard blood pressure control:

• improves recovery and ability to live independently;

• improves health-related quality of life;

• improves survival;

• increases or reduces bleeding in the brain;

• decreases the duration of days in hospital;

• affects the frequency of other unwanted effects.

What did we do?

We searched for clinical studies (in which people are assigned to groups by chance) that included adults with ischemic stroke who had received reperfusion treatment with a clot-busting medicine, mechanical thrombectomy, or both, and compared more intensive blood pressure targets (less than 160 mmHg) with standard targets (less than 180 mmHg) shortly after reperfusion.

We combined and summarized the results of the studies. We also considered how the studies were conducted, the number of participants included, and the similarities and differences between the studies to judge how much confidence we can have in the results.

What did we find?

We found nine studies, involving 4381 adults with ischemic stroke who had reperfusion treatment. Most participants were treated in hospitals in high-income countries, with a smaller number in hospitals in middle-income countries. None of the studies took place in low-income countries. In most studies, the intensive strategy aimed to keep systolic blood pressure (the top number in a blood pressure reading) below 140 mmHg, while the standard strategy aimed to keep it below about 180 mmHg.

Main results

• Intensive blood pressure lowering makes little or no difference to the chance of living independently about three months after the stroke, compared with standard blood pressure control.

• Intensive blood pressure lowering probably makes little or no difference to health-related quality of life.

• Differences in brain bleeding (including severe bleeding that worsens symptoms) between intensive and standard blood pressure strategies were small and uncertain.

• Intensive blood pressure lowering probably increases the risk of death and probably increases the chance of a poor recovery after stroke.

Overall, the results suggest that lowering blood pressure to very low levels after reperfusion is unlikely to provide additional benefit and might cause harm in some people, compared with keeping blood pressure within more standard targets.

What are the limitations of the evidence?

The evidence has several important limitations, such as the relatively small numbers of people in some studies, which reduce the reliability of the results. Additionally, most participants were treated in specialized centers in middle and high-income countries; there is no evidence from low-income countries. Some groups of people may be under-represented (for example, older adults with many other health problems), and many studies did not provide detailed results separately for women and men.

Because of these limitations, the results should be interpreted with caution. More research is needed to determine whether specific groups of people might benefit from different blood pressure targets.

How up to date is this evidence?

The evidence in this review is up to date as of March 2025.

Key messages

• Compared to placebo (a 'dummy' treatment with no active ingredient), weight-loss medicines may reduce body mass index (which measures whether someone's weight is in a healthy range) and weight slightly. Weight-loss medicines probably make little to no difference to the number of people experiencing unwanted or harmful events.

• We do not know if weight-loss medicines plus lifestyle changes are more effective than lifestyle changes alone, or whether they lead to any unwanted events.

• To better understand the benefits and risks, we need more studies that follow young people for at least one year and measure their physical and mental well-being.

What is obesity?

Obesity means having too much body fat. Obesity in children (0 to 9 years) and adolescents (10 to 19 years) increases the risk of several diseases in the short term and adulthood. Obesity also affects how young people feel about themselves and interact with others, often resulting in low self-esteem, anxiety, depression, and social difficulties.

How is obesity treated?

Obesity is a complex problem with many factors. Children and adolescents with obesity may benefit from lifestyle changes, such as improvements in diet and exercise, counseling, and behavioral support. Weight-loss medicines are typically only considered when lifestyle changes have not worked.

Several medicines are used for obesity in children and adolescents, including newer medicines called GLP-1 agonists, better known by brand names such as Wegovy and Ozempic. We considered the effects of all weight-loss medicines together.

What did we want to find out?

We wanted to learn if weight-loss medicines are better than a placebo (a 'dummy' treatment with no active ingredient) or lifestyle changes at:

• reducing body mass index (BMI) — a number calculated from a person's weight and height that indicates whether their weight is in a healthy range, expressed as kilograms per meter squared (kg/m2);

• reducing weight (kg);

• improving well-being.

We also wanted to find out if these medicines lead to:

• unwanted events;

• people leaving the studies early due to unwanted events;

• obesity-related health problems, including high blood sugar, high blood pressure, and unhealthy blood fat levels.

What did we do?

We searched for studies that compared weight-loss medicines with placebo, no treatment, another weight-loss medicine, or lifestyle changes in children and adolescents, and lasted at least three months. We compared and summarized the results of the studies and rated our confidence in the evidence based on factors such as study methods and the number of young people involved.

What did we find?

We included 37 studies with 4218 young people. Twenty-eight studies involved adolescents only. Eight included children and adolescents. One study planned to include children as well as adolescents, but did not mention their ages. Thirty-one studies compared weight-loss medicines to placebo. Six compared weight-loss medicines plus lifestyle changes (changes to diet, physical activity, or both) to lifestyle changes alone.

Main results

Compared to placebo, weight-loss medicines:

• may reduce BMI by an average of 1.8 kg/m2, based on evidence from 25 studies with 3091 young people. This is a small reduction, and may not meaningfully improve young people's health;

• may reduce weight by an average of 5.47 kg (20 studies, 2380 people);

• probably make little to no difference to young people's well-being (4 studies, 741 people);

• probably make little to no difference to the number of people experiencing unwanted events, which was high in both groups: 872 of 1000 people given weight-loss medicines and 846 of 1000 given placebo experience unwanted events (8 studies, 1877 people);

• may make little to no difference to the number of people leaving studies early due to unwanted events: 36 people per 1000 given medicine leave early, compared to 24 people per 1000 given placebo (13 studies, 2213 people);

• may make little to no difference to obesity-related health problems (1 study, 51 people).

There are important differences in the reductions in BMI and weight for different weight-loss medicines.

We are very uncertain about the effects of weight-loss medicines plus lifestyle changes compared to lifestyle changes alone on:

• BMI change;

• weight change;

• the number of people leaving studies early due to unwanted events.

No studies in this comparison measured young people's well-being, the number who experienced unwanted events, and obesity-related health problems.

What are the limitations of the evidence?

For both comparisons, our confidence in the evidence for most outcomes was reduced because the studies used methods likely to introduce errors in their results.

For weight-loss medicines versus placebo, our confidence in the evidence for BMI and weight was also reduced because the studies' results varied. We had little confidence in the evidence about the number of people who left studies early due to unwanted events because it was based on only a few cases. We had only moderate confidence in the evidence for well-being because there were too few studies to be certain about the results.

How up to date is this evidence?

The evidence is current to July 2023. In June 2025, we checked the status of ongoing studies, and updated the results accordingly.

Key messages

• Several medications can help rheumatoid arthritis symptoms after failure of a biologic or targeted synthetic (b/ts) disease‐modifying antirheumatic medicine (DMARD) (a type of medication that stops or slows the disease process).

• We are unsure how the medications compare with placebo (dummy treatment) or with each other in terms of unwanted effects.

• We need more studies that directly compare drugs to get a better idea of which is best. Treatment decisions should consider individual health factors and preferences.

What is rheumatoid arthritis?

Rheumatoid arthritis is a disease where the body’s immune system, which normally fights infections, mistakenly attacks the joints. This causes joint pain, swelling, stiffness, and disability. Rheumatoid arthritis is a long-term condition, but treatments can help manage symptoms and slow the damage.

How is rheumatoid arthritis treated?

Rheumatoid arthritis is typically treated with disease-modifying antirheumatic drugs (DMARDs), which work to control the inflammation responsible for symptoms and also prevent joint damage. There are various kinds of DMARDS. Targeted synthetic (ts) DMARDs work by targeting specific cells or proteins. Biologic (b) DMARDs also target specific cells or proteins, but are made from material that comes from living organisms. Lastly, conventional synthetic DMARDs affect the whole immune system rather than targeting specific parts of it.

What did we want to find out?

We wanted to explore the effectiveness of DMARDs in adults with rheumatoid arthritis for whom treatment with a b/ts DMARD didn't work. This is a 'living' systematic review, meaning that it will be updated regularly. As this field of medicine is always growing and changing, we want to have the most up-to-date evidence reflected in this review.

What did we do?

We searched for studies that tested the effectiveness of DMARDs in adults with rheumatoid arthritis for whom treatment with one b/ts DMARD didn't work.

We looked at how well the available DMARDs reduced the symptoms of rheumatoid arthritis, including pain, joint swelling, and issues with joint movement. We measured this using the American College of Rheumatology 50% response criteria, or ACR50 response. An ACR50 response means the rheumatoid arthritis medication is effective and symptoms have improved by at least half. We also looked at any unwanted effect that might cause a person to stop using the medication.

We compared and summarised the results of the studies and rated our confidence in the evidence based on factors such as study methods and sizes.

What did we find?

We found 19 studies that involved 4779 people with rheumatoid arthritis who had tried a tumour necrosis factor inhibitor (a type of bDMARD) in the past, but it did not improve their symptoms. Most people in these studies were women, with an average age of 49 to 58 years, who had been living with rheumatoid arthritis for an average of 6 to 14 years. The studies were conducted worldwide, including in North America, Europe, and Asia. Nine studies were funded by pharmaceutical companies, while the other 10 were funded by national research grants.

Several medications reduced the symptoms of rheumatoid arthritis in patients who have failed a b/ts DMARD:

• a different tumour necrosis factor inhibitor (2 studies; 343 people);

• sarilumab (1 study, 365 people);

• tocilizumab 4 mg/kg (1 study, 319 people);

• tocilizumab 8 mg/kg (1 study, 328 people);

• intravenous (into a vein) abatacept (2 studies, 665 people);

• rituximab (1 study, 499 people);

• upadacitinib (1 study, 333 people);

• tofacitinib (1 study, 263 people);

• a higher dose of baricitinib (1 study, 353 people).

We are unsure how the medications compare with placebo (dummy treatment) or with each other in terms of unwanted effects, as the evidence for this outcome was generally of low quality.

What are the limitations of the evidence?

Our confidence in the evidence varied from high to very low. In cases where our confidence was limited, this was because there were not enough studies to be certain about the results. When we searched for studies to include in the review, we found some gaps in the available evidence. Specifically, we found that there were fewer studies that directly compared medications with each other and more studies that compared medications to a placebo. More studies that compare drugs directly would better help us understand which drug is best.

How up to date is this evidence?

The evidence is current to 28 November 2025. As this is a living review, we will be updating this research yearly to ensure our results reflect new evidence.

Key messages

• Compared to no physical activity, any type of physical activity may help children with obesity to reduce their BMI slightly (the 'body mass index' is used to check if someone is a healthy weight) and may lead to considerably more minor unwanted events, but we are very uncertain about these results.

• We do not know if physical activity has any effect on the weight, body fat percentage, well-being, and blood sugar levels of children with obesity compared to no physical activity or another type or amount of physical activity. We also do not know if physical activity has any effect on the number of serious unwanted effects.

• We need larger, longer-lasting, and better-quality studies to be more certain about the effects of physical activity on children with obesity.

What is obesity?

Obesity is when a person has too much body fat. Worldwide, obesity amongst children (aged 0 to 9 years) has increased substantially since 1975. More than 300 million children and adolescents (aged 10 to 19 years) may be living with obesity by 2050.

Children with obesity are more likely to develop health problems such as heart disease, stroke, type 2 diabetes, some cancers, liver disease, asthma, joint and muscle problems, and infections later in life. Childhood obesity is also linked to higher rates of anxiety and depression.

Weight loss is the best way to reduce the health risks caused by obesity. In children, this is usually done through lifestyle changes such as doing more physical activity (e.g. running games or interactive video games that encourage movement (sometimes called 'exergaming')) and eating healthier foods.

What did we want to find out?

We wanted to find out if, compared to no physical activity, physical activity helps children with obesity:

• reduce their BMI (the 'body mass index' indicates if someone is a healthy weight; their weight is divided by their height multiplied by itself, expressed as kg/m2) or BMI z-score (how someone's BMI compares to BMIs of others of the same age and sex);

• lose weight;

• improve their well-being;

• reduce body fat percentage or improve how body fat is distributed in the body; and

• lower their blood sugar levels.

We also wanted to find out if one type or amount of physical activity was better than another type or amount at producing these benefits, and if the physical activity programmes led to any unwanted effects.

What did we do?

We searched for studies that compared participating in physical activity to:

• no physical activity;

• another type or amount of physical activity.

The children needed to do the physical activities for at least 12 weeks. We compared and summarised the results of the studies and rated our confidence in the evidence, based on factors such as study methods and number of participants.

What did we find?

We found four studies including 517 children. The studies were small: the smallest included only 59 children and the largest, 222. Just under half of the children involved were girls (46%). Their average age was 9 years.

In three studies, the children did the physical activities for 12 or 13 weeks; the fourth study's physical activity programme lasted 32 weeks. The studies took place in three countries: one study each in Brazil and Iran, and two studies in the USA. In the American studies, 73% of the children were Black. The other studies did not describe the children's ethnicities.

• All four studies compared physical activity to no physical activity.

• One study also compared one type of physical activity (exergaming) to another type (swimming pool exercises).

• One study also compared doing 20 minutes versus 40 minutes of the same physical activities.

Main resultsPhysical activity versus no physical activity

Compared to no physical activity, children with obesity who participated in any physical activity programme may have:

• slightly reduced their BMI: the BMI of children in the active groups was on average 1.52 kg/m2 lower (2 studies, 118 children);

• experienced more minor unwanted effects: children in the active groups may have been about 3.5 times more likely to experience minor unwanted effects (1 study, 222 children).

However, we are very uncertain about these results.

We do not know if physical activity makes any difference to children's:

• BMI z-score;

• weight;

• well-being;

• body fat percentage or how fat is distributed in the body;

• blood sugar levels;

• risk of experiencing serious unwanted events.

One type or amount of physical activity versus another type or amount

We are very uncertain about the effect of one type or amount of physical activity compared with another type or amount on any of the points we were interested in.

What are the limitations of the evidence?

We have limited confidence in the evidence because:

• there were few studies and they were small;

• the types of physical activities varied;

• the studies' reporting of their methods and findings was poor;

• three out of four studies did not report all their results.

How up to date is this evidence?

This evidence is current to December 2025.

Key messages

• Medicines that help the body's immune system recognise and attack cancer cells can shrink the cancer and likely slow its return in people aged 65 years and older having surgery for the most common type of lung cancer. However, it probably does not increase how long people live overall.

• We do not have enough evidence yet to know if its benefits outweigh the harms, and no study assessed well-being.

What is resectable non-small cell lung cancer?

About 85% of people with lung cancer have what is called non-small cell lung cancer (NSCLC). It is most often found in older adults, with an average age at diagnosis of 71. When the cancer has not spread out of the lung, and if the person is fit enough to have surgery, the tumour can be removed. However, the cancer can come back after treatment, and between 30% and 55% of people eventually die from the disease. To reduce the risk of cancer coming back, different treatments can be used together.

How is resectable non-small cell lung cancer treated?

In recent years, lung cancer treatment has been transformed with the discovery of medicines that help the body's immune system recognise and attack cancer cells (immunotherapy). Immunotherapy works together with drugs that kill cancer cells (chemotherapy). When given before surgery, after surgery, or both, immunotherapy improves the chances of curing the cancer and helps delay its return.

What did we want to find out?

The ageing process affects the immune system, reducing its capacity to fight cancer: this is called immunosenescence. The elderly population is a very diverse group, and each person needs to be carefully assessed because age alone is not enough to base the best treatment decisions on. We wanted to find out if, despite immunosenescence, immunotherapy is effective in people aged 65 years and older, to cure lung cancer and prolong life without unwanted treatment effects.

What did we do?

We searched for high-quality studies that tested immunotherapy, either alone or in combination with chemotherapy, given before surgery, after surgery, or both, in people aged 65 and older. We summarised their results and assessed how confident we can be in the evidence, based on criteria such as how the studies were done and how many people were included.

What did we find?

For people aged 65 and older, we found results from 11 studies involving 3152 people with NSCLC. The studies compared immunotherapy alone or in combination with chemotherapy to a placebo (a sham or ‘dummy’ treatment) or no treatment, with or without chemotherapy. Treatments were given either only before surgery, only after surgery, or both before and after surgery.

Immunotherapy reduces the amount of cancer found at surgery and probably delays the cancer from coming back. Immunotherapy probably makes little or no difference to overall survival. There was not enough evidence to determine its effect on potential harms, and no study assessed well-being.

What are the limitations of the evidence?

Our confidence in most of the evidence is moderate because of some concerns about how the studies were conducted. More complete and longer-term data are needed to better estimate how much a person's lifespan might be extended. We have very limited confidence in the assessment of unwanted effects, because only one study provided data about these.

How up to date is this evidence?

The evidence is up to date to 3 July 2025.

Key messages

• Overall, we did not find sufficient evidence to show that medicines that block interleukin-1 (a protein involved in immune responses) are effective treatments for people with COVID-19, or whether they cause unwanted effects.

• We found 16 studies with unpublished results. We will update this review when new data are available.

• In future, we need high-quality studies to evaluate medicines that block interleukin-1 to treat COVID-19.

What is interleukin-1 and what is its role in COVID-19?

Interleukin-1 (IL-1) is a type of protein called a cytokine, which helps to regulate the body’s immune system. In particular, IL-1 triggers inflammation to help fight infection. In COVID-19, as the immune system fights the virus, the lungs and airways become inflamed, causing breathing difficulties. However, in some people, the immune system can over-react (called a ‘cytokine storm’) and produce dangerously high levels of inflammation and tissue damage. This can lead to severe breathing difficulties, organ failure and death.

What are interleukin-1 ‘blockers’?

IL-1 blockers are medicines that stop IL-1 from working by blocking signals from IL-1 to other parts of the immune system. This reduces inflammation and may help the immune system to fight COVID-19. In turn, this may reduce the need for breathing support with a ventilator (a machine that breathes for a patient) and reduce the number of deaths from COVID-19. Three IL-1 blockers are available: anakinra, canakinumab and rilonacept.

What did we want to find out?

We wanted to know if IL-1 blockers are effective treatments for people with COVID-19, compared with standard care alone or with placebo (a dummy treatment that appears identical to the medicine being tested but without any active medicine). We were particularly interested in the effects of IL-1 blockers on:

• whether people’s symptoms got better or worse;

• how many people died; and

• any unwanted effects and serious unwanted effects.

What did we do?

We searched for studies that assessed the effects of IL-1 blockers to treat people with COVID-19 compared with standard care alone or with placebo. People in the studies could have suspected or confirmed COVID-19 of any severity (mild, moderate or severe), and be any age or sex.

We compared and summarised the results of the studies and rated our confidence in the evidence, based on factors such as study methods and sizes.

What did we find?

We found six studies with 2132 people. Four studies assessed anakinra (1633 people) and two assessed canakinumab (499 people). People in the studies were aged between 58 and 68 years old on average, and the majority were men. All the people in the studies were in hospital, mainly with moderate to critical COVID-19. The studies varied in size, from 45 to 2253 people. At the start of the studies, 67% to 100% of people were receiving oxygen, and 0% to 33% were on a ventilator.

We also found 16 studies that have not yet published their results.

Anakinra compared to usual care and placebo to treat people with COVID-19

• Anakinra probably results in little or no improvement in COVID-19 symptoms (defined as improvement on a clinical scale or discharge from hospital) at 28 days after treatment (three studies, 837 people) but we do not know if it makes a difference at 60 days (one study, 115 people).

• We do not know if anakinra makes a difference to the number of deaths at 28 days after treatment (two studies, 722 people) or at 60 days (four studies, 1633 people).

• Anakinra probably results in little or no increase in any unwanted effects at 28 days after treatment, but we are not sure about its effect on serious unwanted effects (two studies, 722 people).

Canakinumab compared to usual care and placebo to treat people with COVID-19

• Canakinumab probably results in little or no improvement in COVID-19 symptoms (defined as improvement on a clinical scale or discharge from hospital) at 28 days after treatment (two studies, 499 people).

• We do not know if canakinumab makes a difference to the number of deaths at 28 days after treatment (two studies, 499 people) or at 60 days (one study, 45 people).

• Canakinumab probably results in little or no increase in any unwanted effects (one study, 454 people), but we are not sure about its effect on serious unwanted effects (two studies, 499 people) at 28 days.

What are the limitations of the evidence?

Our confidence in the evidence is limited for several reasons. All the people in the studies were hospitalised, but some were more seriously ill than others - some studies only included people on a ventilator. Usual care also differed between studies, and studies measured and reported their results using different methods.

How up to date is this evidence?

The evidence is up to date to 5 November 2021.