Abstract
[Purpose] This study aimed to determine the accuracy of combining general movements assessment and head magnetic resonance imaging in predicting cerebral palsy in preterm and low birth weight infants. [Participants and Methods] This retrospective observational study analyzed clinical routine data of patients admitted to the neonatal intensive care unit of a university hospital between 2010 and 2017. The study included 154 very low birth weight infants (mean gestational age 28.8 ± 3.1 weeks, birth weight 1003.7 ± 307.4 g). The infants were classified into a cerebral palsy group or a non-cerebral palsy group. We examined whether combining general movements assessment and head magnetic resonance imaging findings were associated with the development of cerebral palsy. Additionally, we calculated the sensitivity and specificity of each assessment. [Results] All assessments were associated with the presence or absence of cerebral palsy. Among them, the highest sensitivity and specificity, 92.7% and 97.9%, respectively, were observed in infants with abnormal head magnetic resonance imaging findings or poor general movements, even if the head magnetic resonance imaging was not classified as abnormal. [Conclusion] The combination of general movements assessment and head magnetic resonance imaging findings is useful for predicting cerebral palsy.
Keywords: General movements, Magnetic resonance imaging, Cerebral palsy
INTRODUCTION
Cerebral palsy is a neurodevelopmental disorder characterized by abnormal motor function and posture1), caused by disturbances in brain development before or after birth. Its incidence is high worldwide and is even higher among very low birth weight (VLBW) infants (birth weight <1,500 g) compared to term infants2, 3). Children with cerebral palsy experience various challenges in daily life, including abnormal muscle tone, impaired motor coordination, and difficulty maintaining posture1, 4). Rehabilitation, therefore, plays a crucial role in cerebral palsy management.
Early diagnosis and intervention are essential for the treatment and rehabilitation of cerebral palsy5). An observational technique known as general movements (GMs), developed by Prechtl et al., has been reported to have high reliability and effectiveness in predicting cerebral palsy6, 7). However, a single evaluation of GMs is not recommended due to its high false-positive rate8, 9). Its predictive accuracy may improve when combined with other diagnostic tools, such as head magnetic resonance imaging (MRI), during neonatal intensive care unit (NICU) admission9).
This study aimed to assess the predictive value of GMs evaluation at the expected date of birth in VLBW infants for the onset of cerebral palsy, both individually and in combination with head MRI findings during NICU admission. This study builds on our previous research on the topic, “Evaluation to predict cerebral palsy in very low-birth weight infants”10). Previous studies have predicted cerebral palsy from individual perinatal factors; however, in the present study, we tested whether the combined use of new assessment methods would improve the accuracy of predicting cerebral palsy.
PARTICIPANTS AND METHODS
This survey was a single-center, retrospective, observational study. The participants were 230 preterm infants weighing less than 1,500 g of birth weight who were admitted to the neonatology department of the first author’s university hospital between April 2010 and March 2017. The inclusion criterion was a birth weight of <1,500 g. The exclusion criteria were: (I) diagnosed chromosome abnormalities, malformation syndromes, or neurological disorders, (II) death, (III) infants who were not evaluated for GMs or head MRI, and (IV) infants who were not followed up as outpatients due to transfer or relocation. Eleven infants were diagnosed with chromosome abnormalities, malformation syndromes, or neurological disorders, one infant had died, eight infants were not evaluated for GMs or head MRI, and 56 infants were not followed up as outpatients due to transfer or relocation. Thus, 154 infants (mean gestational age: 28.8 ± 3.1 weeks, birth weight: 1003.7 ± 307.4 g) were included in the analysis. The diagnosis of cerebral palsy was made by a physician by the age of two, and the infants were classified into either the cerebral palsy (CP) group (n=14) or the non-cerebral palsy (non-CP) group (n=140). Table 1 presents the characteristics of the participants.
Table 1. Characteristics of the CP group and non-CP group.
| All | CP group | non-CP group | |
| n=154 | n=14 | n=140 | |
| Multiple pregnancy, n (%) | 27 (17.5) | 2 (14.3) | 25 (17.9) |
| Male, n (%) | 86 (55.8) | 7 (50.0) | 79 (56.4) |
| Birth weight, g* | 1,003.7 ± 307.4 | 814.6 ± 371.0 | 1,022.6 ± 295.3 |
| Gestational age, weeks* | 28.8 ± 3.1 | 26.1 ± 3.0 | 29.1 ± 2.9 |
| Small for gestational age, n (%) | 64 (41.6) | 3 (21.4) | 61 (43.6) |
| Head circumference at birth, cm* | 24.9 ± 2.8 | 23.1 ± 3.4 | 25.1 ± 2.7 |
| 1-minute Apgar score† | 5.0 ± 2.4 | 3.5 ± 2.3 | 5.1 ± 2.4 |
| 5-minute Apgar score† | 6.7 ± 1.9 | 5.6 ± 1.8 | 6.8 ± 1.9 |
| Intraventricular hemorrhage† | |||
| Normal, n (%) | 122 (79.2) | 4 (28.6) | 118 (84.3) |
| Grade I–II, n (%) | 23 (14.9) | 6 (42.9) | 17 (12.1) |
| Grade III–IV, n (%) | 5 (3.2) | 2 (14.3) | 3 (2.1) |
| Periventricular echo densities | |||
| Normal, n (%) | 92 (59.7) | 4 (28.6) | 88 (62.9) |
| Grade I, n (%) | 42 (27.3) | 6 (42.9) | 36 (25.7) |
| Grade II–III, n (%) | 15 (9.7) | 1 (7.1) | 14 (10.0) |
| Periventricular leukomalacia, n (%)‡ | 14 (10.0) | 3 (21.4) | 0 (0) |
| Retinopathy of prematurity† | |||
| Normal, n (%) | 72 (46.8) | 1 (7.1) | 71 (50.7) |
| Grade I–II, n (%) | 51 (33.1) | 6 (42.9) | 45 (32.1) |
| Grade III–V, n (%) | 29 (18.8) | 5 (35.7) | 24 (17.1) |
| Laser photocoagulation, n (%)‡ | 43 (27.9) | 9 (64.3) | 34 (24.3) |
| Days of mechanical ventilation, days* | 19.4 ± 24.2 | 38.2 ± 32.8 | 17.5 ± 22.5 |
| Respiratory distress syndrome, n (%) | 73 (47.4) | 9 (64.3) | 64 (45.7) |
| Chronic lung disease 28, n (%) | 93 (60.4) | 11 (78.6) | 82 (58.6) |
| Home oxygen therapy, n (%)‡ | 4 (2.6) | 2 (14.3) | 2 (1.4) |
| Sepsis, n (%)‡ | 37 (24.0) | 7 (50.0) | 30 (21.4) |
| Symptomatic patent ductus arteriosus, n (%)‡ | 18 (11.7) | 3 (21.4) | 15 (10.7) |
| Patent ductus arteriosus ligation, n (%) | 4 (2.6) | 1 (7.1) | 3 (2.1) |
| Late-onset circulatory collapse, n (%) | 21 (13.6) | 2 (14.3) | 19 (13.6) |
| Length of hospital stay, days* | 111.4 ± 57.6 | 159.1 ± 85.7 | 106.6 ± 52.1 |
*p<0.05: unpaired t-tests, †p<0.05: Mann–Whitney U tests, ‡p<0.05: Fisher’s exact tests.
CP: cerebral palsy; non-CP: non-cerebral palsy.
GMs and head MRI results were investigated retrospectively from medical records. The assessment of GMs was performed around the expected date of birth, with a mean corrected age at assessment of 38.7 ± 2.1 weeks. Observations were conducted in a private room or quiet area with the infant in a back-lying position, wearing minimal clothing, during alert wakefulness, and actively moving. GMs were recorded using a video camera positioned diagonally above the infant to capture 1 to 5 minutes of spontaneous movements. Based on the criteria of Hadders-Algra et al.11), GMs were categorized into four groups: “normal optimal (NO)”, “normal suboptimal (NS)”, “mildly abnormal (MA)”, and “definitely abnormal (DA)”. The assessments were conducted by two physiotherapists who were regularly involved in neonatal rehabilitation. In cases where their assessments differed, the video was re-evaluated, and a consensus was reached through discussion.
Almost all infants underwent head MRI under light sedation during admission at the university hospital. MRI evaluations were performed by a physician and assessed for abnormal white matter signals, reduction in white matter volume, cystic changes, myelination/thinning of the corpus callosum, and ventricular dilation. The MRI findings were classified as either abnormal or not abnormal12). In this study, four assessment methods were used to predict the onset of cerebral palsy: (1) MA & DA GMs, (2) DA GMs, (3) abnormal MRI, and (4) a combination of abnormal MRI and DA GMs. The combination of abnormal MRI and DA GMs included infants with abnormal head MRI findings and those whose MRI results were not classified as abnormal but who exhibited DA GMs.
Unpaired t-tests, Mann–Whitney U tests, and Fisher’s exact tests were used to compare characteristics between groups. Fisher’s exact tests were conducted to determine whether abnormalities in each assessment method were associated with a cerebral palsy diagnosis. The diagnostic accuracy of the assessment methods was determined by calculating sensitivity, specificity, positive predictive value, and negative predictive value. The statistical significance level was set at <5%, and statistical analyses were performed using the Statistical Package for the Social Sciences.
This study was approved by our institution’s university hospital ethics committee. Additionally, an opt-out system was implemented for the release of research information, ensuring participants’ right to refuse participation.
RESULTS
Assessments of GMs were as follows: 0 NO, 0 NS, 10 MA, and 4 DA in the CP group (n=14), and 2 NO, 51 NS, 80 MA, and 7 DA in the non-CP group (n=140). In the CP group, 13 infants had abnormal head MRI findings, while in the non-CP group, 31 infants had abnormal head MRI findings. Table 2 presents the sensitivity and specificity of GMs and MRI in predicting cerebral palsy. MA & DA GMs, DA GMs, abnormal MRI, and the combination of abnormal MRI and DA GMs were all significantly associated with cerebral palsy (p=0.002, p=0.01, p<0.001, p<0.001, respectively).
Table 2. Sensitivity and specificity of GMs and head MRI in predicting cerebral palsy (%).
| Sensitivity | Specificity | PPV | NPV | |
| MA+DA GMs* | 100 | 37.9 | 13.9 | 60.9 |
| DA GMs* | 28.6 | 95.0 | 36.4 | 93.0 |
| Abnormal MRI* | 92.9 | 77.9 | 29.5 | 99.1 |
| Abnormal MRI+DA GMs* | 92.9 | 97.9 | 81.3 | 99.3 |
*p<0.05: Fisher’s exact tests.
GMs: general movements; MRI: magnetic resonance imaging; MA: mildly abnormal; DA: definitely abnormal; PPV: positive predictive value; NPV: negative predictive value.
DISCUSSION
In this study, we assessed the predictive value of GMs in VLBW infants for the onset of cerebral palsy, both independently and in combination with head MRI. We found that the presence of abnormal GMs at the expected date of birth and abnormal head MRI findings were significant predictors of cerebral palsy. Furthermore, the combination of abnormal head MRI findings and DA GMs demonstrated high sensitivity and specificity.
A systematic review on the prognostic prediction of cerebral palsy using GMs reported a sensitivity of 75–100% and a specificity of 40–48%13). The sensitivity and specificity of GMs for predicting cerebral palsy in preterm infants were reported as 33% and 64%, respectively14). Additionally, the sensitivity and specificity of head MRI for predicting cerebral palsy in VLBW infants, when evaluated before NICU discharge, were 71% and 91%, respectively15). Previous studies have found that while both GMs and head MRI are useful in predicting neurological prognosis in VLBW infants, their individual sensitivity and specificity are not sufficiently high. However, this study demonstrated that the combined evaluation of GMs and head MRI could enhance the prediction of cerebral palsy in VLBW children. Specifically, the combination of abnormal MRI and DA GMs showed a high sensitivity of 92.7% and specificity of 97.9% for cerebral palsy development.
Cerebral palsy can occur even in infants with normal head MRI findings. Approximately 10% of children diagnosed with cerebral palsy have normal head MRIs16). In such cases, cerebral palsy is often non-spastic, ataxic, and mild in severity17). GMs, on the other hand, are useful in predicting a broader range of prognoses, including cognitive dysfunction in addition to cerebral palsy18). Therefore, combining head MRI with GMs may allow for the identification of children with cerebral palsy, including those with mild severity, who may not show abnormalities on MRI.
The findings of this study align with previous research that recommends a combined evaluation rather than a single assessment in predicting cerebral palsy prognosis. These results are significant as they suggest that NICU staff can consider early diagnosis and intervention for cerebral palsy in VLBW infants. Additionally, rehabilitation staff may be able to provide more careful developmental follow-up for VLBW infants suspected of having cerebral palsy based on GMs and head MRI findings.
A limitation of this study is that cerebral palsy diagnoses were made by the age of 2 years; therefore, some infants in the study may later be diagnosed with cerebral palsy or other developmental disabilities. Future multicenter studies should be conducted to track the long-term prognosis of VLBW infants.
In conclusion, the presence of abnormal GMs and abnormal MRI findings was predictive of cerebral palsy. Furthermore, combining GMs evaluation with MRI findings improved predictive accuracy.
Conflicts of interest
The authors declare no conflict of interest.
REFERENCES
- 1.Patel DR, Neelakantan M, Pandher K, et al. : Cerebral palsy in children: a clinical overview. Transl Pediatr, 2020, 9: S125–S135. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Linsell L, Malouf R, Morris J, et al. : Prognostic factors for cerebral palsy and motor impairment in children born very preterm or very low birthweight: a systematic review. Dev Med Child Neurol, 2016, 58: 554–569. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Pascal A, Govaert P, Oostra A, et al. : Neurodevelopmental outcome in very preterm and very-low-birthweight infants born over the past decade: a meta-analytic review. Dev Med Child Neurol, 2018, 60: 342–355. [DOI] [PubMed] [Google Scholar]
- 4.Østensjø S, Carlberg EB, Vøllestad NK: Everyday functioning in young children with cerebral palsy: functional skills, caregiver assistance, and modifications of the environment. Dev Med Child Neurol, 2003, 45: 603–612. [DOI] [PubMed] [Google Scholar]
- 5.Hadders-Algra M, Boxum AG, Hielkema T, et al. : Effect of early intervention in infants at very high risk of cerebral palsy: a systematic review. Dev Med Child Neurol, 2017, 59: 246–258. [DOI] [PubMed] [Google Scholar]
- 6.Ferrari F, Cioni G, Prechtl HF: Qualitative changes of general movements in preterm infants with brain lesions. Early Hum Dev, 1990, 23: 193–231. [DOI] [PubMed] [Google Scholar]
- 7.Prechtl HF, Einspieler C, Cioni G, et al. : An early marker for neurological deficits after perinatal brain lesions. Lancet, 1997, 349: 1361–1363. [DOI] [PubMed] [Google Scholar]
- 8.Kwong AK, Fitzgerald TL, Doyle LW, et al. : Predictive validity of spontaneous early infant movement for later cerebral palsy: a systematic review. Dev Med Child Neurol, 2018, 60: 480–489. [DOI] [PubMed] [Google Scholar]
- 9.Bosanquet M, Copeland L, Ware R, et al. : A systematic review of tests to predict cerebral palsy in young children. Dev Med Child Neurol, 2013, 55: 418–426. [DOI] [PubMed] [Google Scholar]
- 10.Uchio Y, Shima N, Nakamura K, et al. : Evaluation to predict cerebral palsy in very low-birth weight infants. J Jpn Acad Health Sci, 2022, 25: 18–24 (in Japanese). [Google Scholar]
- 11.Hadders-Algra M, Mavinkurve-Groothuis AM, Groen SE, et al. : Quality of general movements and the development of minor neurological dysfunction at toddler and school age. Clin Rehabil, 2004, 18: 287–299. [DOI] [PubMed] [Google Scholar]
- 12.Woodward LJ, Anderson PJ, Austin NC, et al. : Neonatal MRI to predict neurodevelopmental outcomes in preterm infants. N Engl J Med, 2006, 355: 685–694. [DOI] [PubMed] [Google Scholar]
- 13.Darsaklis V, Snider LM, Majnemer A, et al. : Predictive validity of Prechtl’s method on the qualitative assessment of general movements: a systematic review of the evidence. Dev Med Child Neurol, 2011, 53: 896–906. [DOI] [PubMed] [Google Scholar]
- 14.John HB, Oommen SP, Swathi TO, et al. : Preterm general movements in prediction of neurodevelopmental disability and cerebral palsy at two years: a prospective cohort study. Indian Pediatr, 2022, 59: 769–773. [PubMed] [Google Scholar]
- 15.Mirmiran M, Barnes PD, Keller K, et al. : Neonatal brain magnetic resonance imaging before discharge is better than serial cranial ultrasound in predicting cerebral palsy in very low birth weight preterm infants. Pediatrics, 2004, 114: 992–998. [DOI] [PubMed] [Google Scholar]
- 16.Leonard JM, Cozens AL, Reid SM, et al. : Should children with cerebral palsy and normal imaging undergo testing for inherited metabolic disorders? Dev Med Child Neurol, 2011, 53: 226–232. [DOI] [PubMed] [Google Scholar]
- 17.Springer A, Dyck Holzinger S, Andersen J, et al. : Profile of children with cerebral palsy spectrum disorder and a normal MRI study. Neurology, 2019, 93: e88–e96. [DOI] [PubMed] [Google Scholar]
- 18.Einspieler C, Bos AF, Libertus ME, et al. : The general movement assessment helps us to identify preterm infants at risk for cognitive dysfunction. Front Psychol, 2016, 7: 406. [DOI] [PMC free article] [PubMed] [Google Scholar]