Descriptive Epidemiology of Anorectal Malformations in Chinese Population

Over a twelve year period one hundred and five neonates (30 females) with an anorectal malformation (ARM) were studied at a tertiary referral centre in Hong Kong. A perineal fistula was the most frequent presentation of ARM in males closely followed by rectourethral fistula. Vestibular fistula was the most common anomaly amongst females. An ARM without a fistula was documented in 11 % of patients and dominated by males. Associations and syndromes were present in nearly 30 % of the group with VACTERL association and Down syndrome at the forefront. Isolated anomalies were most frequently seen in the renal tract (23 %), the CNS (20 %) and the heart (18 %). A quarter of the cohort was small for gestational age with syndromes/associations featuring in half of this subgroup. A death rate of 9.5 % was recorded, of which all except one had an underlying syndrome or association. In addition a pre-existing heart condition constituted the most significant factor affecting the mortality in 50 % of this group.


INTRODUCTION
ARM manifests itself as an embryological anomaly of the hindgut between the 4th and 8th week of gestation.The pathogenesis is largely unknown, but it is thought the earliest defect leading to the development of an ARM is a deficiency in the dorsal component of the cloacal membrane; the severity of the ARM is related to the extent of this deficit [1].This study sets out to analyse the various clinical presentations and its associations / syndromes in a Chinese population.The charts and computer records of all patients with ARM were reviewed analyzing demographics, the details of the ARM, and co-morbidities ranging from syndromes to single organ system anomalies.Furthermore causes for mortality and long-term survival were scrutinized.

All
The ARMs were categorized according to the Krickenbeck classification with the necessary separation between male and female gender [2,3].All babies underwent surveillance imaging studies: chest and abdominal radiographs, two-dimensional echocardiography, abdominal and renal sonography.A routine ultrasound of the lumbosacral spine was supplemented by an magneto resonance imaging (MRI) scan, if required.
Suspected chromosomal anomalies were confirmed by chromosomal analysis.A VACTERL association, first described in 1965, was considered, if at least three of the possible seven criteria were identified [4][5][6].
Isolated anomalies were categorised by organ system, e. g., the gastrointestinal tract.Anomalies affecting one organ system were registered once for the individual patient even if an infant had two or more anomalies within the same organ system, e.g., an atrial septum defect (ASD) alongside a ventricular defect (VSD).
The often quoted genitourinary system was divided into renal tract and genital area.Placing genital anomalies together with renal tract aberrations would have distorted the analysis, as genital pathologies (e.g., hypospadias or undescended testis) are mainly seen in boys.
The statistical analysis was conducted with a Fisher's exact test for differences in categorical data.SPSS software (SPSS.22)was utilised.The P-value (P) was set at .05 for significance.

RESULTS
One hundred and five neonates (30 females) were admitted with an ARM.The females were born at a median gestation of 38 (range 31-42) weeks with a median weight of 2.56 (range 1.12 -3.69) kg.Eleven female neonates were small for gestational age (SGA) (Figure 1) .The males were born at the same median gestation of 38 (range 30 -42) weeks with a median weight of 2.84 (range 1.2 -4) kg; in two boys the weight was not recorded.Fifteen male babies were small for gestational age (Figure 1).In total, 26 (24.8 %) were SGA.
The ARM in both male and female subjects were subclassified in Table 1.Twelve infants presented with an imperforated anus without a fistula.Two had an ARM which could not be verified in detail before their death.
Syndromes and associations accounted for 30 cases (28.6 %) within the cohort.Sixteen (15 %) had a VACTERL association, chromosomal anomalies were seen in 10 (9.5 %), two cases (1.9 %) were part of a genetic syndrome, one presented with the Mayer-Rokitansky-Küster-Hauser (MRKH) syndrome (Müllerian agenesis) and one patient had multiple anomalies not (yet) linked to a specific syndromic complex.
The differences between the gender are illustrated in Table 2.Although it appears the smaller female subgroup is more often affected by associations and syndromes; the difference is not statistically significant (P = .054).Regarding isolated anomalies in organ systems the renal tract system is most often affected [N=24 (23 %)], closely followed by central nervous system (CNS) and vertebra column anomalies [N= 21 (20 %)].In this subgroup girls were more often affected (P= .032).A tethered cord was seen in only 3 ( 2.9 %) patients (Table 3).
Cardiac anomalies were noted in 19 infants (18 %) with another preponderance for females in this system (P=.003).Further anomalies in the gastrointestinal, genital, musculoskeletal and craniofacial systems affect 3.8 to 16 % of the cohort with no gender preference (Table 4).Esophageal atresia / tracheo-esophageal fistula (EA/TEF) and duodenal atresia are the most common gastrointestinal conditions associated with ARM.
Overall 10 (9.5 %) patients (5 females) died; five were neonates with SGA.Although the female group was overall smaller, the difference did not reach statistical significance (P = .116).Associated anomalies were seen in 9 of the deaths: four VACTERL associations, one Down syndrome, one Edwards syndrome, one Treacher Collins syndrome, one Turner variant, and one with multiple, non-syndromic anomalies (Table 5).In four cases death was precipitated by a pre-existing heart condition and cardiac failure was a significant factor in the death of a fifth patient.
Three death could be attributed to an (associated) surgical condition: one had a rectal perforation of his ARM prior to stoma formation leading to sepsis, one died of the sepsis following perforation of an associated malrotation volvulus and the third had an aspiration pneumonia in conjunction with a tracheomalacia post EA/TEF repair.
Only two individuals died of factors unrelated to their underlying anomalies: one had total parenteral nutrition (TPN) related liver failure and one suffered with pneumonia.

DISCUSSION
The first attempt to classify ARMs dates back to 1835 when Amussat described several groups of ARMs [7].This study is based on a consensus classification system agreed upon at the Krickenberg conference [2].It was developed because of dissatisfaction with previous classification systems; the relatively arbitrary definitions of 'high', 'intermediate' and 'low' ARMs were difficult to translate into clinical practice [8].
The Krickenberg classification reflects the anatomical findings of ARMs, as they were identified through the experience with the posterior sagittal anorectoplasty and the radiological studies of the relationship between the renal tract, the genital area and the hindgut structures (e.g.fistula) [3].
The classification identifies major clinical groups of ARM whilst making allowances for rare conditions and regional variants [2].
In this study perineal fistula is overall the most common ARM affecting more than a third of the group.Separating the cohort along gender lines the perianal fistula continues to be the most frequent ARM in boys followed by rectourethral fistula, but in girls the most dominant ARM is a rectovestibular fistula (Table 1).These results are confirmed by the findings of Nah et al. [9] , but partially contradicted by Levitt's and Chen's assertion that rectourethral fistula dominates the ARM in boys [10,11].
ARM without a fistula is considered a relatively uncommon condition.Large scale studies from the US and Japan identified imperforated anus without a fistula ranging from just 4 % to 5 % respectively [12,13].However in our cohort this condition was twice more common [N=12 (11.4 %)] with only 2 affected females.One-third had Down syndrome.Nah et al. reported a similar rate of 13 % patients with ARM without a fistula in a Singaporean population, albeit they did not specify the frequency of Down syndrome within this subgroup [9].Likewise Cho et al. identified an ARM without a fistula at a rate of 15.5 % within their cohort of predominantly Caucasian infants [14].
Moore asserts in his monograph on the epidemiology of ARM a near equal ratio between male and female ARMs; we were unable to confirm this finding [15].We found a clear male preponderance resulting in a male to female ratio of 2.5 : 1.Our observation was confirmed by other authors with male to female ratios ranging from 1.46 : 1 to 2.4 : 1 [9,13,14,16].
The gestational age was a median of 38 weeks for both boys and girls being similar to other studies [17].We noted a quarter of the neonates were SGA.Just over half of these small-for-date babies had a syndrome or VACTERL association; this finding was confirmed in the series by Stoll et al. [16].
Syndromes and associations accounted for 28.6 % (30 infants) in this cohort.Chromosomal anomalies alone were seen in 10 (9.5 %); this is in line with other series [9,14,18].Down syndrome is the most common chromosomal anomaly associated with ARM present in 6.7 % of the cohort and similar to the 2 to 9 % observed by other authors [11,18].
The three most frequently affected organ systems were the renal system (23 %) followed by the CNS (20 %) and the heart (18 %).
Within the subgroup of CNS and vertebra anomalies a tethered cord was seen in only 3 patients (2.8 %).Nah et al. reported 20 infants (20 %) with tethered cord in their cohort [9].We have no explanation for this discrepancy in particular as all our patients underwent spinal ultrasound screening followed by an MRI scan, if required.Notably, Stoll et al. identified just one tethered cord in their series of 174 ARMs, and other authors reported no tethered cord at all [11,16].
Cardiac anomalies [N=19 (18%)] were only considered, if clinically relevant, e.g., a spontaneously closing patent ductus arteriosus (PDA) was not included.Godse et al. reported a similar rate of cardiac anomalies [17].We noted a predilection for girls with cardiac anomalies ( p = .003);a finding not reported previously.
The mortality in our cohort is overall 9.5 % (10 individuals); the seemingly higher proportion of deaths in the smaller female group (5/30,16.7 %) is not statistically significant (P = .116).Half of all mortality cases were small for gestational age (SGA) babies.
Other series report similar or higher mortality rates: Godse et al. had a 12.6 % mortality in a Caucasian cohort and an Indian series reported a 16 % death rate [17,19].
Looking at the underlying causes it is noteworthy all deaths except one were seen in babies with a syndrome or an association (Table 5).Five infants (50%) neonates with an ARM admitted to this tertiary referral centre between 1. Jan 2002 and 31.Dec 2013 were included into the group.The local clinical research ethical committee approved the study (Ref.No.: CRE-2013.148).

Figure 1 :
Figure 1: Percentile weight distribution of infants with ARM; all babies with weight less than 10 th percentile considered SGA.(Note that no infants had intrauterine growth retardation).

Table 3 :
Structural anomalies in the renal tract, the CNS and the cardiovascular system

Table 4 :
Structural anomalies in the gastrointestinal, genital, musculoskeletal and craniofacial area