Structured clinical question

In infants with delayed umbilical cord separation (POPULATION), does investigating all for possible leucocyte adhesion deficiency (LAD) (INTERVENTION), compared with only those with signs of immune impairment (COMPARISON) act as an effective screening tool for LAD (OUTCOME).

Search

A Cochrane Library Search for produced no relevant reviews. PubMed searches for ‘umbilical cord separation AND time’ and ‘leucocyte adhesion deficiency’ yielded 404 (five relevant) and 724 (eight relevant) results (table 1). Non-English language, single case reports and papers relating to LAD in other species/animal models were excluded.

Commentary

To establish if there is benefit to investigating infants with delayed cord separation, we must define ‘delay’.

Cord separation occurs at the demarcation zone following infiltration of polymorphonuclear leucocytes and collection formation between the abdominal wall and the base of the drying stump.1 2

We reviewed five cohort studies (2023 infants). In the largest study (Oudesluys-Murphy3) of 911 neonates, separation times ranged from 1 to 29 with a mean of 7.4 days (SD±3.3). Mean separation times in other studies4–7 ranged from 7.4 to 13.9 days (longest reported 45 days). No studies followed infants beyond cord separation time, but none had severe/recurrent infection or LAD at study closure. Novack et al 5 reported 38% of patients with cord separation time >2 weeks and Wilson et al 6 10% at >3 weeks. Although delayed separation is defined as occurring later than 2–3 weeks, a significant proportion of healthy infants will experience delayed separation.

A screening test aims to identify disease risk in apparently healthy individuals, allowing early intervention.8 An effective screening test requires a clinically important disease. LAD is a rare (1:100 000), hereditary, autosomal recessive primary immunodeficiency, with three subtypes. LAD causes defects in the leucocyte adhesion cascade (a four-step process involving rolling, firm arrest, postarrest modifications and transmigration into tissues9) preventing neutrophil migration to the site of tissue infection (online supplemental figure 1).

LAD-I and LAD-III are fatal in infancy/childhood without haematopoietic stem cell transplantation.

Establishing a case for screening requires a screening test. Patients with LAD demonstrate peripheral blood neutrophilia (neutrophils are unable to exit the circulation). However, tissue and other extravascular spaces are profoundly neutropaenic, producing the clinical manifestations of LAD.10 A full blood count (the proposed intervention) is a first-line investigation with definitive diagnosis by flow cytometry. Early diagnosis would allow for monitoring and earlier treatment.

First described in 1979,11 the most common LAD, LAD-I, is caused by a loss of function mutations in the ITGB2 gene (encoding for CD18, subunit of integrin12), causing a defect in firm arrest of leucocytes to endothelium.13

Clinical features include recurrent severe bacterial infections, particularly of soft tissue, with absence of pus formation, omphalitis and delayed umbilical cord separation in the neonatal period and impaired wound healing.14

LAD-II is milder, rarer and caused by mutations SLC35C1 gene encoding a fucose-specific GDP transporter,15 affecting the rolling phase of the leucocyte adhesion cascade.16 Patients present with persistent periodontitis, developmental delay, intellectual impairment, dysmorphic features, poor growth, Bombay type blood group and metabolic consequences of impaired fucose metabolism.10

LAD-III/LAD-I variant is rarer and less severe then LAD-I and is caused by a FERMT3 gene mutation, leading to defects in the Kindlin-3 component of β-integrins.17 In addition to other features of LADs, LAD-III has associated severe bleeding tendencies.18

If we propose to investigate all infants with delayed cord separation for possible LAD, we must establish the sensitivity and specificity of the initial screening question.

In Almarza Novoa et al,10 the most comprehensive publication, umbilical complications (delayed cord separation/omphalitis) were reported in 84% of severely affected and 58% of moderately affected patients. The other case reports19–22 describe delayed umbilical cord separation in just 60%–78.9% of patients, with omphalitis in 12.5%–92%. Therefore, delayed cord separation is not pathopneumonic of LAD-I.

Fewer studies relating to LAD-II/LAD-III met inclusion criteria. There are no reported cases of delayed umbilical cord separation in LAD-II8 12. In LAD-III delayed cord separation is present in <50% of patients, suggesting a weaker association.19 ²⁵

Of the 273 patients included in all studies (for whom data on umbilical complications was available), delayed cord separation and/or omphalitits (not differentiated in the largest study)10 was present in 181 patients. Presence of delayed cord separation as an initial screening test would have a sensitivity of 66.3% (online supplemental table 1).

Because we do not know the number of patients with delayed cord separation with LAD, we must extrapolate from existing data. LAD is rare (1:100 000), so a hypothetical population of 100 000 infants would contain one case.

Extrapolating the proportion of infants with cord separation >2 weeks, 38% (Novack4) n=38 000 and >3 weeks 10% (Wilson5) n=10 000 would give a specificity of 61.9%–62% at 2 weeks and 89.9%–90% at 3 weeks, depending on if affected case was true positive (online supplemental tables 2.1–2.4). Although ‘presence of delayed cord separation’ gives a relatively high specificity at 3 weeks, the high rate of delayed separation time in the ‘normal’ population and the relatively low sensitivity make it a poor initial screening tool for LAD.

In conclusion, cord separation time in healthy infants is variable. Delayed cord separation is common but not universal in LAD-I patients, is weakly associated with LAD-III and is apparently absent in LAD-II. Therefore, while delayed cord separation should raise index of suspicion of a LAD in patients with recurrent/severe soft tissue infections and poor wound healing, there is unlikely to be any benefit in investigating all otherwise healthy infants with delayed cord separation. In addition, delayed cord separation should not be a prerequisite for investigation for LAD, as do so would risk delayed diagnosis.