Key messages
• Currently, we do not know when to stop giving caffeine to preterm babies.
• If we stop the caffeine too early, babies may have brief, sometimes repetitive, episodes when they have lower levels of oxygen in their blood.
• We need to do more studies to find out the best time to stop giving caffeine to preterm babies.
Why is caffeine given to premature babies?
Caffeine is a stimulant used to prevent and treat apnea – when breathing repeatedly stops and starts – in preterm infants. Caffeine also seems to reduce episodes of intermittent hypoxia - rapid moments when the blood oxygen levels drop. Caffeine reduces the risk of the baby needing mechanical ventilation (a therapy that assists or replaces spontaneous breathing) to provide enough oxygen supplementation. It improves lung development, reduces the risk of developing chronic lung damage, and improves the quality of life of preterm babies.
When is caffeine given?
Caffeine is given to preterm infants, but there is currently no clear guidance on precisely when to start it. Some studies have found the best benefits when started within the first three days of life.
There is also no clear guidance on when to stop giving caffeine. Some investigators stop it at a defined post-menstrual age, which is the infants' age, calculated from the first day of their mother's last menstrual period. Others stop when the babies are without symptoms for five days or more.
What did we want to find out?
We wanted to find out the best time to stop caffeine. We measured this as the need to resume therapy, the need for mechanical ventilation, the number of episodes of apnea or intermittent hypoxia, survival, brain development in infancy and childhood, and the duration and cost of hospitalization.
What did we do?
We searched for studies that compared:
• the cessation of caffeine administration at different post-menstrual ages;
• the cessation of caffeine at a definite post-menstrual age or after a symptom-free period;
• the cessation of caffeine after different periods free from symptoms (five days or more).
We compared and summarized the results of the studies and rated our confidence in the evidence based on factors, such as study methods and sample sizes.
What did we find?
We included three studies with 392 preterm infants, with a gestational age between 26 and 33 weeks. They were randomly divided into two groups that either stopped or continued giving caffeine.
Two studies compared stopping caffeine at a definite post-menstrual age versus when the babies were free of symptoms. One study compared stopping caffeine at two different post-menstrual ages. We did not find any studies that compared stopping caffeine after different symptom-free periods.
Stopping caffeine at different post-menstrual ages
• Stopping caffeine early (before 35 weeks of post-menstrual age) may increase the number of intermittent hypoxemia episodes in the seven days after stopping, compared to stopping later.
• Stopping caffeine early (before 35 weeks of post-menstrual age) may make little to no difference in the number of deaths before discharge from hospital compared to stopping later.
• The evidence is very uncertain about whether there is any difference in unwanted effects between stopping caffeine early and late.
Stopping caffeine when the babies are free of symptoms (early) or at a definite post-menstrual age (later)
• Stopping caffeine early may make little to no difference in the number of infants with apnea in the seven days after stopping, compared to stopping later.
• Stopping caffeine early probably results in more infants with episodes of intermittent hypoxia in the seven days after stopping, compared to stopping later.
• Stopping caffeine early may make little to no difference in the number of deaths before discharge, compared to stopping later.
• The evidence is very uncertain about whether there is any difference in unwanted effects between stopping caffeine early and late.
Three studies are ongoing.
What are the limitations of the evidence?
Our confidence in the evidence is limited because the number of babies studied for each outcome was small. We did not find any studies that compared stopping caffeine during different symptom-free periods. Finally, the evidence did not cover all the outcomes in which we were interested.
How up to date is this evidence?
The evidence is up-to-date until August 2023.
There may be little or no difference in the incidence of all-cause mortality and apnea in infants who were randomized to later discontinuation of caffeine treatment. However, the number of infants with at least one episode of IH was probably reduced with later cessation.
No data were found to evaluate the benefits and harms of later caffeine discontinuation for: restarting caffeine therapy, intubation within one week of treatment discontinuation, or need for non‐invasive respiratory support within one week of treatment discontinuation.
Further studies are needed to evaluate the short-term and long-term effects of different caffeine cessation strategies in premature infants.
Apnea and intermittent hypoxemia (IH) are common developmental disorders in infants born earlier than 37 weeks' gestation. Caffeine administration has been shown to lower the incidence of these disorders in preterm infants. Cessation of caffeine treatment is based on different post-menstrual ages (PMA) and resolution of symptoms. There is uncertainty about the best timing for caffeine discontinuation.
To evaluate the effects of early versus late discontinuation of caffeine administration in preterm infants.
We searched CENTRAL, PubMed, Embase, and three trial registries in August 2023; we applied no date limits. We checked the references of included studies and related systematic reviews.
We included randomized controlled trials (RCTs) in preterm infants born earlier than 37 weeks' gestation, up to a PMA of 44 weeks and 0 days, who received caffeine for any indication for at least seven days. We compared three different strategies for caffeine cessation: 1. at different PMAs, 2. before or after five days without symptoms, and 3. at a predetermined PMA versus at the resolution of symptoms.
We used standard Cochrane methods. Primary outcomes were: restarting caffeine therapy, intubation within one week of treatment discontinuation, and the need for non‐invasive respiratory support within one week of treatment discontinuation. Secondary outcomes were: number of episodes of apnea in the seven days after treatment discontinuation, number of infants with at least one episode of apnea in the seven days after treatment discontinuation, number of episodes of intermittent hypoxemia (IH) within seven days of treatment discontinuation, number of infants with at least one episode of IH in the seven days after of treatment discontinuation, all-cause mortality prior to hospital discharge, major neurodevelopmental disability, number of days of respiratory support after treatment discontinuation, duration of hospital stay, and cost of neonatal care. We used GRADE to assess the certainty of evidence for each outcome.
We included three RCTs (392 preterm infants).
Discontinuation of caffeine at PMA less than 35 weeks' gestation versus PMA equal to or longer than 35 weeks' gestation
This comparison included one single completed RCT with 98 premature infants with a gestational age between 25 + 0 and 32 + 0 weeks at birth. All infants had discontinued caffeine treatment for five days at randomization. The infants received either an oral loading dose of caffeine citrate (20 mg/kg) at randomization followed by oral maintenance dosage (6 mg/kg/day) until 40 weeks PMA, or usual care (controls), during which caffeine was stopped before 37 weeks PMA.
Early cessation of caffeine administration in preterm infants at PMA less than 35 weeks' gestation may result in an increase in the number of IH episodes in the seven days after discontinuation of treatment, compared to prolonged caffeine treatment beyond 35 weeks' gestation (mean difference [MD] 4.80, 95% confidence interval [CI] 2.21 to 7.39; 1 RCT, 98 infants; low-certainty evidence). Early cessation may result in little to no difference in all-cause mortality prior to hospital discharge compared to late discontinuation after 35 weeks PMA (risk ratio [RR] not estimable; 98 infants; low-certainty evidence).
No data were available for the following outcomes: restarting caffeine therapy, intubation within one week of treatment discontinuation, need for non‐invasive respiratory support within one week of treatment discontinuation, number of episodes of apnea, number of infants with at least one episode of apnea in the seven days after discontinuation of treatment, or number of infants with at least one episode of IH in the seven days after discontinuation of treatment.
Discontinuation based on PMA versus resolution of symptoms
This comparison included two RCTs with a total of 294 preterm infants.
Discontinuing caffeine at the resolution of symptoms compared to discontinuing treatment at a predetermined PMA may result in little to no difference in all-cause mortality prior to hospital discharge (RR 1.00, 95% CI 0.14 to 7.03; 2 studies, 294 participants; low-certainty evidence), or in the number of infants with at least one episode of apnea within the seven days after discontinuing treatment (RR 0.60, 95% CI 0.31 to 1.18; 2 studies; 294 infants; low-certainty evidence). Discontinuing caffeine based on the resolution of symptoms probably results in more infants with IH in the seven days after discontinuation of treatment (RR 0.38, 95% CI 0.20 to 0.75; 1 study; 174 participants; moderate-certainty evidence).
No data were available for the following outcomes: restarting caffeine therapy, intubation within one week of treatment discontinuation, need for non‐invasive respiratory support within one week of treatment discontinuation, or number of episodes of IH in the seven days after treatment discontinuation.
Adverse effects
In the Rhein 2014 study, five of the infants randomized to caffeine had the caffeine treatment discontinued at the discretion of the clinical team, because of tachycardia.
The Pradhap 2023 study reported adverse events, including recurrence of apnea of prematurity (15% in the short and 13% in the regular course caffeine therapy group), varying severities of bronchopulmonary dysplasia, hyperglycemia, extrauterine growth restriction, retinopathy of prematurity requiring laser treatment, feeding intolerance, osteopenia, and tachycardia, with no significant differences between the groups.
The Prakash 2021 study reported that adverse effects of caffeine therapy for apnea of prematurity included tachycardia, feeding intolerance, and potential neurodevelopmental impacts, though most were mild and transient.
We identified three ongoing studies.