OBJECTIVE: We determined the frequency and estimated effective radiation dose (E) from conventional diagnostic radiographs for infants who had birth weight of ≤1500 g (very low birth weight [VLBW] infants) and were treated in a NICU.
METHODS: Entrance skin doses were experimentally measured for all standard weight-dependent exposure settings. For each radiograph in the radiologic file, the exposed area on the film was measured manually. Together with clinical data obtained from the Vermont Oxford Network, medical charts, and radiologic files, we estimated E. E values per radiograph and per child were compared with recommended reference values and annual natural background radiation (NBR). We used reference data to estimate the risk for radiation-induced cancers.
RESULTS: Of 212 VLBW infants, 194 required at least 1 conventional radiograph. Measured entrance skin dose varied between 11.8 and 15.0 μGy. Calculated E received was 16 microsievert (μSv; median) per radiograph and 71.5 μSv (median) per infant for the whole stay. Infants with birth weight ≤750 g, length of stay ≥16 weeks, congenital malformations, or oxygen dependence for ≥36 weeks were at risk for high numbers of radiographs and high radiation dose. Compared with the annual NBR, the median of 4 radiographs per infant contributes 12 days of NBR. We estimated that only 1 of 60000 NICU-treated VLBW infants will develop a fatal malignancy up to 15 years of age.
CONCLUSIONS: We found that NICU-treated VLBW infants had low radiation exposure compared with the annual NBR.
- radiation exposure
- radiation dose
- effective dose
- preterm neonates
- very low birth weight infants
- extremely low birth weight infants
- neonatal intensive care unit
- conventional radiographs
Very low birth weight (VLBW) infants are at risk for multiple medical and surgical problems and therefore may require multiple radiologic examinations during a short period of time. Radiation risk strongly depends on the age at which exposure occurs; therefore, preterm neonates are at highest risk for radiation-induced malignancies1,2 because of the potentially large number of radiographs performed together with neonates' greater potential life span3 and the longer opportunity for occurrence of radiation-induced malignancies. It is therefore important to monitor and minimize radiation doses to preterm infants, especially to VLBW infants. The International Commission on Radiologic Protection (ICRP) has defined the concept of effective dose (E),4 which has become an accepted method for estimating dose and associated risks in diagnostic radiology.5 There have been few international comparisons of E for preterm,6–8 VLBW9, and extremely low birth weight (ELBW) infants.8,10 To our knowledge, this is the first German study on the estimation of E in a large number of VLBW and ELBW infants. The aims of the study were to determine the frequency of conventional radiographic procedures, to estimate E per radiograph and per child during the entire stay in the NICU, to compare E per child with recommended values and with background radiation, and to estimate the increased risks for radiation-induced cancer.
The NICU of the Division of Neonatology Charité, Campus Mitte, is a level 3, 31-bed NICU with 1700 to 1800 deliveries per year. All infants who had birth weight ≤1500 g and were admitted to the NICU between January 1, 1997, and December 31, 1999, were eligible for the survey. Patients were followed from the day of NICU admission until their discharge home or to another hospital.
Clinical patient data were obtained from the Vermont Oxford Network (VON), a voluntary group of health professionals who are committed to improving the quality and safety of medical care for neonates. Since 1995, the NICU of the Charité, Campus Mitte, has been an active member of the VON (www.vtoxford.org11). All live-born infants who have birth weight of 401 to 1500 g and are born in or transferred to the participating hospital before 28 days' postnatal age are eligible for the database. All data in this study were validated and compared with data from >250 voluntary member NICUs. Additional clinical information came from medical records, including charts and daily flow sheets. Data for thoracic diameters relative to postnatal age were obtained from published standard ultrasound data for the fetal thorax (12–41 weeks).12 Fetal and postnatal growth have been assumed to be equal, although this is not applicable for all patients. Radiographic reports were reviewed retrospectively, and the date, type (chest, abdomen, combined chest/abdomen, or skeletal), frequency, and exposed area of each conventional radiograph were recorded. All data were double checked with the electronic information system (MEDORA [GE Medical Systems, Germany, Munich]). Imaging results other than those of conventional radiographs were classified and recorded but not analyzed regarding E.
Radiographic Device and Technical Setting
All radiographs were taken with the patient in the supine position by using a portable radiograph machine Mobilett B unit (Siemens-Elema, Solna, Sweden), used exclusively in the NICU. Radiograph machine settings came from a weight-dependent table used by the radiographers. A film screen system with a sensitivity class of 400, an additional filtration of 1.0 mm of aluminum and 2 mm of copper, and an adequate protective shielding was used. No anti-scatter grid or automatic exposure control was applied, and a focus film distance of 100 cm was chosen. These settings were consistent with the recommendations of the Commission of the European Communities (CEC) for a standard radiologic technique for neonates.13
Measurement of Entrance Skin Dose and Calculation of Dose-Area Product
Entrance skin dose (ESD; free in air without backscatter) was measured with the diagnostic dose meter DIADOS (PTW, Freiburg, Germany). Because measurement of ESD for each patient during each radiographic procedure was technically and organizationally impractical, measurements were performed experimentally, without a neonate, under NICU conditions (incubator setting and DIADOS placed at the height of the estimated skin entrance level). The ESD of 10 repeated measurements for each of the 4 weight-dependent technical exposure settings varied between mean ± SD 11.80 ± 0.10 μGy and 14.80 ± 0.13 μGy. For additional estimation of E, according to our conservative design, we used an ESD of 15 μGy. The field size of the film was calculated by manual measurements of length and width for each of the 1553 radiographs to the nearest 0.5 mm. Dose-area product was calculated by multiplying the ESD at the patient entrance surface with the field size (exposed area on film). For calculations of E (see next section), these results were linked with the specific body weight of the neonate at the time of each examination.
Estimation of E
For all additional evaluations of E and risk estimation, the 194 infants with at least 1 conventional procedure were included. This corresponds to our conservative design of E to overestimate rather than underestimate. Generally, calculations of E are time-consuming and difficult, especially under daily clinical conditions. A practical method that uses the energy imparted together with conversion factors obtained from Monte Carlo calculations has been developed for various radiograph projections14 and extended to pediatric patients.15,16 Imparted energy can be determined from measurements of dose-area product, and, in combination with tabulated values, E can be estimated by using the methods already described when the following parameters are available for each radiograph: body thickness (cm), body weight (g), projection view, tube voltage (kV), half-value layer (mmAl), ESD (μGy), and exposed film area (cm2). All results of E were expressed in microsievert (μSv).
Estimation of Risk
Recommended reference values for newborns exist for ESD but not for E13; therefore, E per radiograph and per child during the whole length of stay (LOS) was compared with the latest dates of annual natural background radiation (NBR) in Germany.17 The risk for radiation-induced cancer has been calculated according to the risk for prenatal exposure of 2.8 to 13 × 10−2/Sv−1 given by the ICRP.4 We chose these data because most radiographs were taken before the neonates reached the 37th week of gestational age (GA). According to our conservative design, to overestimate rather than underestimate the risk, we used a maximal risk factor of 13 × 10−2/Sv−1 in our calculations.4
Data analysis was performed by using SPSS 9.0 and 10.0 (SPSS, Chicago, IL). Results are reported as percentages and averages. Medians were used in non-normal data distribution. Nonparametric Mann-Whitney U tests were used to compare skewed variables between 2 groups, and Kruskal-Wallis statistics were used to analyze skewed variables among ≥ groups. Results of experimental measurements (normal data distribution) were expressed as mean ± SD. For metric data, a linear regression model was used (R-Quadrat). Box and whisker plots according to Tukey were used for graphic visualization. Statistical significance was set at P < .05.
Patient Demographics and Medical Conditions
Of the 212 VLBW infants (median birth weight: 1100 g [range: 445–1500 g]; median GA: 29.5 weeks [range: 24–36 weeks]) who were born between January 1, 1997, and December 31, 1999, 194 required at least 1 conventional radiograph. Overall patient demographics and medical conditions for these neonates, as well as for 4 birth weight categories (≤750 g, 751–1000 g, 1001–1250 g, and 1251–1500 g according to the VON database) are summarized in Table 1. Seventeen neonates died in hospital at a median of 16 days of life (range: 2–204 days); 11 (65%) lived >10 days. These infants required 287 conventional radiographs (median: 9; range: 1–62), or 19% of the total number. Infants who died in the delivery room (n = 8) were excluded. By the 28th day of life, 68.3% of all conventional films were taken, and the median discharge, including discharge to other hospitals, was on the 38th day of life.
Type, Frequency, and Timing of Radiographs
A total of 1553 radiographic procedures were needed for various medical conditions. Fifty (3%) were special radiographic procedures (8 computed tomography scans, 2 micturating cystourethrographies, 30 gastrointestinal tract contrast radiographies, 8 fluoroscopies, and 2 angiographies); these were not further analyzed in this study. More than two thirds of the 1503 conventional radiographs were chest radiographs (67%), followed by combined thoracic/abdominal radiographs (22%), abdominal radiographs (7%), and skeletal radiographs (1%). Eighteen (8.5%) neonates did not require any radiologic examination. Only 3 of these had a birth weight <1000 g (minimum: 850 g); their median GA was 30 weeks. None of these 18 patients required mechanical ventilation or surfactant application. During the first days of life, 6 of these 18 received supplemental oxygen and 5 received continuous positive airway pressure. The remaining 194 neonates received at least 1 (median: 4; range: 0–62) conventional radiograph, with 16% (n = 33) requiring 1 and 32% (n = 67) requiring 2 to 4 radiographic films. In contrast, 8% of the VLBW infants required >20 films, and 4% required >30 films. Altogether, 56% of the study population had 0 to 4 conventional radiographs (Fig 1). The maximum of 62 films was required for an ELBW neonate (GA: 28 weeks; birth weight: 515 g) with severe complications (respiratory distress syndrome, chronic lung disease, patent ductus arteriosus, necrotizing enterocolitis, and sepsis). This patient died on day 169 of life. By the end of the first week of life, 37.1% (n = 557) of the radiograph films had been requested: 13.6% (n = 205) on the first day of life, 27.6% (n = 415) through the third day, and 33.1% (n = 498) through the fifth day. A total of 68.3% of the radiograph films were requested through the 28th day of life.
The frequency of radiographic procedures depended on various conditions, including birth weight (Table 2), GA, and LOS. Neonates with a birth weight <751 g required a median of 11 radiographs, whereas those with a birth weight >1000 g required a median of 2 radiographs (P < .001). Significantly more radiographs were requested for neonates with lower GA (P < .001) and longer LOS (P < .001).
Estimation of E
Median E was 16 μSv per radiograph and 71.5 μSv per infant (range: 8.5–1424.0 μSv) for the entire stay in hospital. Values for special body regions are shown in Table 3. Regression analyses (E = 17.86 * number of radiographs − 3.67; regression coefficient of R2 = 0.95) show that, for each conventional radiograph, the E increased 18 μSv.
E and Medical Conditions
Variables such as birth weight (P < .001), GA (P < .001), LOS (P < .001), and oxygen dependence correlated significantly with E (Figs 2 and 3). Multiple regression analysis showed that birth weight, oxygen administration, LOS, and malformations were the most significant factors influencing actual E levels (P < .001 for all; Table 4), whereas higher birth weight correlated with a significant decrease in E level. A linear regression model of E and field size showed a positive, linearly significant relation (P < .001; regression equation E [μSv] = 0.06 * exposed area [cm2]; ie, for each 1 cm2, the E increased 0.06 μSv).
Of all radiographs taken in the first year of life, ∼60% are thoracic radiographs and 10% are abdominal radiographs.18 Similar distributions of conventional radiographs have been reported for preterm infants and neonates,19,20 for VLBW infants,9 and for ELBW infants.10 We found that 4% of infants required >30 radiographs, lower than the 7% previously reported.8 One study, of 55 VLBW infants, found that the infants received an average of 9.1 (median: 5.0) radiographs,9 whereas another study found that 25% of neonates (birth weight <2500 g) required only 1 radiograph.21 Within our cohort of 212 neonates (birth weight ≤1500 g), 8.5% required no radiographs and 15% needed only 1. The 36 ELBW infants in our study needed a median of 11 radiographic procedures, lower than the medians of 31,10 23,22 and 268 previously reported. It is remarkable that Arad et al23 reported a median of 7 (0–77) for chest and 3 (0–61) for abdominal radiographs in ELBW infants with no change in frequency since 1987 (Table 5). The CEC has published a diagnostic reference value for ESD (multiplication of ESD by the scatter factor21,24) in chest radiographs (anteroposterior) in newborns of 80 μGy, and an achievable ESD of 30 μGy13,25 on the basis of research in pediatric radiology units throughout the European Union.26–29 All technical and procedural requirements specified in the European Guidelines of the CEC13 were met in this study. All of the maximal ESD levels resulted in values clearly below the achievable ESD of 30 μGy,20 indicating that actual dose levels below the reference values are achievable by applying the CEC recommendations. We were able to measure the exposed area (field size) on each of 1441 available radiographs. In contrast, 62 (4.1%) radiographs were unavailable, but their type was documented in the medical records. For these lost films, we used the median of the measured values of field size for each type of radiograph. Table 6 shows a comparison of our field size measurements with other published values. Our measured field sizes were smaller for almost all types of radiograph. The use of aperture openings that are too large results not only in limited contrast or resolution but also in higher radiation exposure of body parts beyond the area of interest.13 Erroneous aperture settings for infants can lead to exposure of many radiosensitive tissues, in the craniocaudal (thyroid, facial bones, and head) and transverse (hollow bones with plenty of red bone marrow) directions. A field enlargement of just 1 cm at the upper and lower field margins (along the longitudinal axis), starting with an original field size of 6 cm2, can lead to a 33% increase in very preterm neonates.25 Although E levels were estimated as described previously,14,16 errors as a result of calculations must be explained. We observed uncertainties between 5% and 11% in establishing the (E/a)I ratio of 8% for the calculation of the imparted energy at 60 kV, in contrast to the Monte Carlo simulation, and of 0.9% to 1.6% in the measurements of ESD, resulting in a variance of actual E of ∼10% to 20%. Because no other method for preterm neonates was available, this method was acceptable and sufficiently accurate. The decisive advantage lies in the ability to take into consideration not only actual radiation parameters but also individual parameters of each patient, including body thickness and body weight. Table 3 shows the E levels of various conventional radiologic procedures in adults, children, and neonates.6,7,10,17,30–32 This comparison is only an approximation, because even for the same radiologic procedure, dose levels are directly related to variations in equipment and depend on age. Nevertheless, conventional radiographs impose a relatively low level of radiation exposure.33 The same is true for the cumulative E levels. The median cumulative E of 71.5 μSv in our study in VLBW infants was nearly 50% lower than the 136 μSv reported for preterm infants of GA <34 weeks.8 The only 2 articles that referred to ELBW neonates (birth weight <750 g) reported cumulative E levels of 720 μSv (n = 25)10 and 497 μSv (n = 35),8 both of which were higher than the 224 μSv (n = 36) reported in our study. Our finding of lower E levels was likely attributable to the lower frequency of radiologic procedures in our study. E levels from special radiographic procedures (eg, computed tomography scans, fluoroscopy) are significantly higher than for conventional radiographs. In this study, only 3% of all imaging methods were not conventional radiographs, affecting 21 of 194 infants (median: 0 [range: 0–12]; 5 neonates required 2, 3 neonates required >2, and the remaining 13 neonates needed only 1 procedure); however, it is clear that, for infants with very complicated clinical courses, the accumulated E will exceed the recommended reference values. In regard to cancer mortality risk, no correlation has been reported to date between intrauterine or postnatal use of diagnostic radiation and an increasing incidence of leukemia or other tumors during childhood, even for preterm infants34–39 (Tables 7 and 8). The ICRP states that cancer mortality risk in childhood is associated with prenatal exposure to 2.8 to 13 × 10−2/Sv−1.4 Tables 7 and 8 show the cancer mortality risk for our cohort of VLBW infants, on the basis of the highest risk level of 13 × 10−2/Sv−1, findings similar to those previously reported.6,7,9,21,40 We also determined risk factors for preterm infants associated with radiographic exposure of specific body regions.7,10,31 Cancer mortality rates after radiation exposure during childhood (age not further classified) have been reported to be 1:1 000 000 for a thorax radiograph (E <50 μSv) and 1:100 000 for an abdominal radiograph (E <100 μSv).30 Another way to demonstrate the probability of damaging effects of radiography is to compare the actual E with a time-equivalent amount of yearly NBR. In 2007, the average yearly NBR was 2.1 mSv,17 varying between 1 and 6 mSv/year, depending on location. The median cumulative E of 72 μSv in our study cohort (range: 8.5–1424.0 μSv) is therefore equivalent to 12 additional NBR-days (range: 1.5 days to 8 months). Exposure to 1 thoracic radiograph or 1 abdominal radiograph is the equivalent of ∼3 NBR-days, and a combined thoracic-abdominal radiograph is equivalent to 4 extra days of NBR. The cumulative E levels in our cohort, including maximum values, all were well below the yearly NBR of 2.1 mSv. The existing natural radiation levels would be reached only after 140 thorax radiographs per year. Even the maximum actual dose of 1.4 mSv to a seriously ill preterm infant remained below 2.1 mSv.
The results presented here indicate that radiation exposure was low compared with the benefits gained, even for the extremely and severely sick preterm neonates in this cohort. It is important, however, to monitor and document radiation exposure carefully. Optimization of technical conditions and reduction of the number of radiographs are important for radiation protection. Our study showed that there is some potential to minimize the number of radiographic examinations, especially of the entire skeletal system (babygram), without endangering patient outcome.
We thank Prof E. Ludwig Grauel, Nikolaos A. Gkanatsios, PhD, and PD Dr sc nat Gerd Schmalisch for support and encouragement.
- Accepted July 9, 2009.
- Address correspondence to Roland R. Wauer, MD, Charité Centrum 17, Division of Neonatology, Charité Campus Mitte, Schumannstrasse 20/21, 10098 Berlin, Germany. E-mail:
Financial Disclosure: The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject:
Preterm neonates are at risk for radiation-induced malignancies because of the potentially large number of radiographs performed. It is important to monitor and minimize radiation doses. Few international studies of effective dose for VLBW and ELBW infants have been published.
What This Study Adds:
Frequency and type of all conventional radiographs performed and effective dose per radiograph and per infant were determined for VLBW and ELBW infants. These data were compared with recommended values and with background radiation. Risks for radiation-induced cancer were estimated.
- ↵United Nations Scientific Committee on the Effects of Atomic Radiation. Sources and effects of ionizing radiation. UNSCEAR 2000 Report to the General Assembly. Available at: www.unscear.org/unscear/en/publications/2000_1.html. Accessed October 10, 2008
- ↵Martin JA, Hsiang-Ching K, Martin JA, et al. Annual summary of vital statistics: 2006. Pediatrics.2008;121 (4):788– 801
- ↵International Commission on Radiological Protection. 1990 Recommendations of the International Commission on Radiological Protection. Oxford, England: Pergamon Press; 1991. ICRP Publication No 60
- ↵Brindhaban A, Al-Khalifah K. Radiation dose to premature infants in neonatal intensive care units in Kuwait. Radiat Prot Dosimetry.2004;111 (3):275– 181
- ↵Donadieu J, Zeghnoun A, Roudier C, et al. Cumulative effective doses delivered by radiographs to preterm infants in a neonatal intensive care unit. Pediatrics.2006;117 (3):882– 888
- ↵Sutton PM, Arthur RJ, Taylor C, Stringer MD. Ionising radiation from diagnostic x rays in very low birthweight babies. Arch Dis Child Fetal Neonatal Ed.1998;78 (3):F227– F229
- ↵Wilson-Costello D, Rao PS, Morrison S, Hack M. Radiation exposure from diagnostic radiographs in extremely low birth weight infants. Pediatrics.1996;97 (3):369– 374
- ↵Horbar JD. The Vermont Oxford Network: evidence-based quality improvement for neonatology. Pediatrics.1999;103 (1 suppl E):350–359
- ↵European Commission. European Guidelines on Quality Criteria for Diagnostic Radiographic Images in Paediatrics. Brussels, Luxembourg: Office for Official Publications of the European Communities; 1996. Report EUR 16261
- ↵Hart D, Jones DG, Wall BF. Estimation of Effective Dose in Diagnostic Radiology From Entrance Surface Dose and Dose-Area Product Measurement. Oxon, United Kingdom: National Radiological Protection Board; 1994. NRPB-R262
- ↵German Ministry for Environment, Nature Conservation and Nuclear Safety. Environmental Radioactivity and Radiation Exposure in the Year 2007: Annual Report on to the Upper and Lower Houses of the German Federal Government. Bonn, Germany: German Ministry for Environment, Nature Conservation and Nuclear Safety; 2008
- ↵Hart D, Jones DG, Wall BF. Coefficients for Estimating Effective Dose From Paediatric X-Ray Examinations. Oxon, United Kingdom: National Radiological Protection Board; 1996. NRPB-R279
- ↵Fletcher EW, Baum JD, Draper G. The risk of diagnostic radiation of the newborn. Br J Radiol.1986;59 (698):165– 170
- ↵McParland BJ, Gorka W, Lee R, Lewall DB, Omojola MF. Radiology in the neonatal intensive care unit: dose reduction and image quality. Br J Radiol.1996;69 (826):929– 937
- ↵Chapple CL, Faulkner K, Hunter EW. Energy imparted to neonates during x-ray examinations in a special care baby unit. Br J Radiol.1994;67 (796):366– 370
- ↵Ono K, Akahane K, Aota T, et al. Neonatal doses from X ray examinations by birth weight in a neonatal intensive care unit. Radiat Prot Dosimetry.2003;103 (2):155– 162
- ↵Hufton AP, Doyle SM, Carty HM. Digital radiography in paediatrics: radiation dose considerations and magnitude of possible dose reduction. Br J Radiol.1998;71 (842):186– 199
- ↵Schneider K. Radiation protection and quality assurance in pediatric radiology. In: Schuster W, Färber D, eds. Radiography: In Infancy & Childhood. Germany: Springer; 1996:19– 31Vol 1 . Berlin
- ↵Horwitz AE, Schweighofer-Berberish K, Schneider K, et al. Selected image quality parameters in a survey using a test phantom in radiological departments and offices in the Federal Republic of Germany. Radiat Prot Dosimetry.1993;49 (1):79– 82
- Schneider K, Fendel H, Bakawski C, et al. Results of a dosimetric study in the European Community on frequent x ray examination in infants. Radiat Prot Dosimetry.1992;43 (1):31– 36
- Schneider K, Kohn MM, Bakowski C, et al. Impact of radiographic imaging criteria on dose and image quality in infants in an EC-wide survey. Radiat Prot Dosimetry.1993;49 (1):73– 76
- ↵Zeiler M, Weisbach M, Weigl A, et al. Patient exposure and radiographic technique in neonatal chest radiography: a survey in Germany. Eur Radiol Suppl.1993;3 :85
- ↵Cook JV. Radiation protection and quality assurance in paediatric radiology. Imaging.2001;13 (4):229– 238
- ↵Jones NF, Palarm TW, Negus IS. Neonatal chest and abdominal radiation dosimetry: a comparison of two radiographic techniques. Br J Radiol.2001;74 (886):920– 925
- ↵Armpilia CI, Fife IA, Croasdale PL. Radiation dose quantities and risk in neonates in a special care baby unit. Br J Radiol.2002;75 (895):590– 595
- Shu XO, Potter JD, Linet MS, et al. Diagnostic X-rays and ultrasound exposure and risk of childhood acute lymphoblastic leukemia by immunophenotype. Cancer Epidemiol Biomarkers Prev.2002;11 (2):177– 185
- Meinert R, Kaletsch U, Kaatsch P, Schuz J, Michaelis J. Associations between childhood cancer and ionizing radiation: results of a population-based case-control study in Germany. Cancer Epidemiol Biomarkers Prev.1999;8 (9):793– 799
- ↵Wraith CM, Martin CJ, Stockdale EJ, McDonald S, Farquhar B. An investigation into techniques for reducing doses from neonatal radiographic examinations. Br J Radiol.1995;68 (814):1074– 1082
- Copyright © 2009 by the American Academy of Pediatrics