This is the first study to report a large series of 57 spontaneously resolved iNIHF cases with a survival rate of 98.2%. Our observation is consistent with the results of recent studies describing a generally favorable prognosis of prenatally resolved hydrops fetalis [11, 16]. In the univariate analysis by Derderian et al, resolution of hydrops prior to delivery portended better survival rates regardless of the NIHF etiology. They observed resolution in 41 NIHF cases and 76% of these fetuses survived . Gilby et al reported antenatal resolution in 17 fetuses, six of them with an idiopathic etiology, and all these fetuses survived . The isolated cases within the idiopathic group are not reported in any of these studies. The isolated subgroup may tend to have a better survival prognosis compared to the entire idiopathic group, which includes cases with unclear fetal structural abnormalities and unexplained suspected diagnoses.
Regarding the data from our long-term follow-up, more than half of the children were found to have anomalies of varying severity. From 28 children 14 had minor or major anomalies, another 6 had mild logopedic, psychosocial or motoric impairment. We suspect that even in the event of complete prenatal resolution, the finding of an iNIHF could be an early indication of abnormal child development. However, major developmental anomalies were diagnosed in only two children.
Haverkamp et al previously reported increased rates of neurological impairment after NIHF and assumed that the neurological long-term outcome depends on the underlying cause of NIHF . They concluded that in particular, survivors with transient benign intrauterine conditions such as idiopathic and lymphatic etiology are at no additional risk for their psychomotor development . This thesis cannot be verified by the results from our long-term follow-up.
Investigating the long-term outcome, we also searched for indications on prenatally undiagnosed conditions that might have been related to the development and resolution of hydrops fetalis. The reports of recurrent skin edema in three children could be indicative of congenital lymphedema. In one of them, we could obtain the results of a postnatal whole genome aCGH analysis. A heterozygote duplication on chromosome 4 was found with unclear clinical significance and unclear correlation to congenital lymphedema. Lately NIHF has been associated to several hereditary lymphedema syndromes; in a systematic review Quinn et al listed six monogenetic lymphatic diseases that are reported more than once in correlation with NIHF .
Clinical and research implications
In the current American guideline by the Society for Maternal-Fetal Medicine (SMFM), NIHF cases are divided into three prognostic categories: first, cases amenable to fetal therapy; second, those with a fatal prognosis; and third, idiopathic cases with a poor but uncertain prognosis . Regarding the third category, based on our data on the isolated subgroup, we would like to point out that survival rates of iNIHF can be favorable, especially if the NIHF completely resolves prenatally.
We cannot answer whether in these survivors, complete prenatal resolution has beneficial effects on the long-term health. A representative comparative group with prenatal persistence of NIHF is lacking in our long-term data: The only case with persistent mild effusion in the last sonographic control at 29 weeks' gestation had a comparable long-term outcome to cases that were completely resolved prenatally.
To identify iNIHF we recommend at least two ultrasound scans, one of them performed after the first trimester, echocardiography, chromosomal and infectious testing, as well as clarifying prenatal diagnostics in case of any suspected disease.
The extent of other reasonable diagnostics for unexplained NIHF has been increasingly discussed lately: in our study period, screening for metabolic and other monogenetic diseases was not performed regularly, but these conditions are now gaining importance in the diagnostic procedure of unexplained NIHF [2, 9, 36]. In the current SMFM guideline, lysosomal enzyme testing is advised in structurally normal fetuses if available .
Recently discussed genetic diagnostics in NIHF include chromosomal microarray analysis (CMA) and whole exome sequencing (WES).
Assuming we had used prenatal CMA for the two cases with severe functional impairment in our study, we hypothesize that in at least one of them (Case 2 in Table 3), this may have prenatally revealed the genetic abnormalities that were found in postnatal CMA. Possibly the abnormality could have been considered in prenatal counselling. However, the clinical impact of this postnatally diagnosed genetic variant is still unknown. Furthermore, current studies suggest that CMA in general has a low diagnostic utility for NIHF [2, 37].
We cannot estimate how many of our cases would have benefited from prenatal WES which was not routinely used in clinical testing in our early study period or even nowadays . Using WES Sparks et al were able to identify a pathogenic or likely pathogenic variant in 29% of NIHF cases unexplained by standard genetic testing . In a meta-analysis of Mone et al the pooled incremental yield of prenatal exome sequencing over chromosomal microarray analysis or karyotyping was 21% in iNIHF .
Monogenetic disorders are increasingly understood as one of the major contributing etiologies to NIHF [2, 5, 8, 9]. Using WES further progress in terms of understanding the pathogenesis and resolution of NIHF can be expected. Simultaneously, it should be borne in mind that there are still ethical aspects to be considered when using WES e.g., regarding the diagnosis of secondary genetic results such as predisposition to cancer or cardiovascular disease .
Strengths and limitations
One strength of our study is the large number of total NIHF cases which allowed the identification of a representative isolated subgroup. The distinct definition of iNIHF minimized uncontrolled variability within this group.
Another strength of our study is that we achieved a comprehensive long-term follow-up with a mean of 10.4 years for 28 survivors of iNIHF. Patient informed consent, based on detailed study information sent and explained to the parents and children enabled us to obtain multidisciplinary long-term data. We consider the records of the children's health status based on parental perception to be particularly valuable for the prenatal counselling situation.
There are certain limitations of this study due to the retrospective study design and large study period of 19 years. This limited the availability of postnatal outcomes: from all reviewed NIHF cases 5.2% had to be excluded due to missing outcome information; for the long-term outcome of the iNIHF survivors we were able to obtain pediatric reports from only 46.7%.
Assuming that parents are more likely to participate when development is perceived as normal , the number of abnormal developments might be underestimated. Therefore, we cannot predict the long-term health of all fetuses with iNIHF. However, the outcome reports provide data on the spectrum of possible anomalies and functional limitations after prenatal iNIHF.
We investigated whether the exclusion criteria in our study led to a non-response bias concerning the iNIHF survival rates. We estimated the maximal possible error that could have been caused by this bias: from the cases excluded due to a lack of postnatal information, three fetuses may have been isolated as they did not show any anomalies in the complete prenatal diagnostic workup. Assuming all of them showed a spontaneous resolution and had a fatal outcome, the baby-take-home rate for the iNIHF group with prenatal resolution would decrease from 98.2% to 93.3%.
Furthermore, our study only provides data on the baby-take-home rate for fetuses that have survived to the second trimester as we required an inconspicuous ultrasound scan in the second trimester for the identification of iNIHF. Therefore, no predictions on the survival prognosis of NIHF at the time of a first trimester examination can be made from our data. Survival rates could be significantly lower if an unremarkable ultrasound in the first trimester was considered sufficient for the diagnosis of iNIHF.
Other limitations of our study collective imply a preselection bias since the study was conducted in a prenatal referral center. Pregnant women from a large population are referred in case of abnormalities or irregular development. We were not able to draw conclusions about incidences in the general population, but this did not prevent the clear identification of iNIHF within the study population.