Journal reference: Rajaraman P, Simpson J, Neta G, et al. Early life exposure to diagnostic radiation and ultrasound scans and risk of childhood cancer: case-control study. BMJ 2011; 342: d472 ^[1]

Link: http://dx.doi.org/10.1136/bmj.d472

Published: 10 February 2011

Evidence cookie says...

The excess risk of childhood cancer from diagnostic imaging in utero and early infancy remains unclear. This study was underpowered for the results to be meaningful. Nevertheless, the results suggest that in utero ultrasound is unlikely to be associated with childhood cancer. Following a precautionary principle, the usual practice of avoiding diagnostic radiation in pregnant women and young children should continue.

Article details:

Study design:

analysis of data from a large multi-centre, case-control study

Study aim:

to examine the association between risk of childhood cancer and exposure to diagnostic radiation and ultrasound scans in utero and in early infancy (age 0-100 days)

Study conclusion:

Although the results for lymphoma need to be replicated, all of the findings indicate possible risks of cancer from radiation at doses lower than those associated with commonly used procedures such as computer tomography scans, suggesting the need for cautious use of diagnostic radiation imaging procedures to the abdomen/pelvis of the mother during pregnancy and in children at very young ages. ^[1]

Critical appraisal:

Methodology (PICO):

Participants: who was studied?

• review of data from the United Kingdom Childhood Cancer Study (UKCCS)
• children aged 14 years or younger in England, Wales and Scotland were eligible
• however, data not abstracted for participants from Scotland, Avon, Dorset and Sommerset so none of these children were included in analysis
• exclusions:
  • Down’s syndrome
  • radiotherapy
• cases:
  • diagnosed between 1992 and 1996 (year of birth 1976 to 1996) with a confirmed malignancy
  • 3133 were interviewed, 87% of eligible cases
  • 2690 had data abstracted from medical records, 60.1% of eligible cases
  • 2656 (60.0%) were included in analysis
• controls:
  • two controls selected from the same population register and matched on sex and date of birth (within one month)
  • if one eligible control refused interview, another eligible control was chosen until two controls participated per case
  • 7619 were interviewed, 64% of eligible controls
  • 4858 had data abstracted from medical records, 40.6% of eligible controls
  • 4854 (40.5%) were included in analysis
• baseline characteristics:
  • mean age: 5.6 years
  • 83.7% had birthweight between 2500 and 4000 g
  • mean maternal age: 27.8 years
  • notably: 1.3% of cases vs 0.1% of the control had Down’s syndrome (not included in analyses)

Intervention: what was the exposure?

• radiation from medical diagnostic procedures while mother pregnant
• ultrasound while mother pregnant
• radiation from medical diagnostic procedures on infant (aged 1-100 days)
• ultrasonographic procedures on infant (aged 1-100 days)

Comparator: what was the control/alternative?

N/A: see details on the control group in participants.

Outcomes: what was measured?

• odds ratios and 95% confidence intervals (CI) to examine the association between cancer and exposure to diagnostic radiation
• odds ratios and 95% CI to examine association between cancer and ultrasound
• final models adjusted for sex of child, age of diagnosis, study region, maternal age and birth weight
• Down’s syndrome excluded from the dataset; association with leukaemia and strong confounding in the dataset

Are the trial results valid?

Internal validity: Wikipedia

Was the defined representative sample of patients assembled at a common point in the course of their disease?

Unclear.

• controls were matched on sex and date of birth (within 1 month) from the same population register
• there were a substantial number of controls (36%) who refused to participate in the study, compared to 13% of cases → there may be issues with selection bias in the controls

Was patient follow up sufficiently long and complete?

N/A.

• cases chosen were those already with the outcome (paediatric cancer) and children up to the age of 14 years were included in the study

Were outcome criteria either objective or applied in a blind fashion?

Yes.

• diagnosis of cancer is objective
• the history of exposure to diagnostic imaging and ultrasound was from a review of medical records which is also objective

If subgroups with different prognoses are identified, did adjustment for important prognostic factors take place?

Yes.

• analyses were adjusted for sex of child, age of diagnosis, study region, maternal age and birth weight
• it is possible that other unidentified confounding exists

What were the results?

Outcomes:

Odds ratio (OR) (95% confidence interval) with in utero exposure:

• all cancers:
  • radiation exposure, OR = 1.14 (95% CI, 0.90 – 1.45)
  • ultrasound exposure, OR = 0.93 (95% CI, 0.79 – 1.09)
  • interpretation: no statistically significant result is detected

Odds ratio (95% confidence interval) in early infancy (0-100 days) exposure:

• all cancers:
  • radiation exposure, OR = 1.19 (95% CI, 0.82 – 1.74)
  • ultrasound exposure, OR = 1.55 (95% CI, 0.89 – 2.70)
  • interpretation: no statistically significant result is detected

Will the results help me care for my patient?

External validity: Wikipedia

Are the participants different to my patient?

• the children in the study were from the United Kingdom
• the study was performed in the 1990s; radiation doses from diagnostic imaging has decreased since that period
• the majority of radiation exposure in utero was in the form of pelvic, chest or abdominal x-ray (88% of procedures)
• only one mother received CT
• the majority of radiation exposure in early infancy was chest x-ray (72% of procedures)

Will this evidence make a clinically important impact on my conclusions about what to offer to tell my patients?

• the evidence is ambiguous as the study lacks sufficient power to make a meaningful interpretation

Study weaknesses (summary)

• the main weaknesses of this study is the lack of power
• effectively all important outcomes have wide confidence intervals that are not statistically significant
• there is a significant possibility of selection bias in the controls
• the conclusions made by the study authors demonstrate biases in interpretation:
  • “Exposure to diagnostic radiography in early infancy was associated with small, non-significant excess risks for all cancers and leukaemia, as well as increased risk of lymphoma on the basis of small numbers. We found no evidence of increased risk of childhood cancer with in utero exposure to ultrasound scans.” (pg. 5, “Discussion”) ^[1]
  • the increased risk of lymphoma is of questionable significance as it was on the basis of only 7 cases
  • moreover, the overall results are questionable: although an OR demonstrating a non-significant excess risk of 1.19 for early infancy radiation exposure, there was a greater magnitude excess risk with an OR of 1.55 for early infancy ultrasound
  • it is probable that random variation, or effects from unadjusted confounding factors are larger than any real effect if it exists

Study strengths (summary)

• use of medical records for assessment of exposure → likely to be more objective then previous studies based on parental memory

Biases and conflicts of interests

• none declared and none seem obvious

Clinical relevance to primary health care

There is a known probable association between in utero exposure to radiation, including diagnostic imaging, and the risk of childhood cancer. This study attempted to assess the risk between in utero and early childhood exposure of radiation and ultrasound.

It is the opinion of the author of this appraisal that this study is insufficiently powered to make a meaningful assessment of the risk. No statistically significant association of increased risk was found. The study does exclude very large excess risks but lower magnitude risks which are clinically significant remain possible. The exception is that in utero ultrasound exposure is unlikely to be associated with excess risk of childhood cancer; OR = 0.93 (95% CI, 0.79 – 1.09).

Australian general practitioners should follow a precautionary principle and avoid the use of diagnostic radiation in pregnant women and young infants where possible. The excess risk (if it exists) from simple x-rays in utero and in early childhood remains unclear.

References

1. Rajaraman P, Simpson J, Neta G, et al. Early life exposure to diagnostic radiation and ultrasound scans and risk of childhood cancer: case-control study. BMJ 2011; 342: d472