Introduction

Kawasaki disease (KD) is one of the commonest childhood vasculitides and usually occurs in children below 5. Recognition of the characteristic signs and symptoms is crucial as the diagnosis remains essentially clinical. However, in cases with incomplete KD, certain laboratory parameters and echocardiographic assessment of coronary arteries may facilitate the diagnosis [1]. KD has lately generated enormous interest among physicians, scientists, and even the lay public due to its association with the novel coronavirus pandemic (SARS-CoV-2/COVID-19) [2]. This is reflected in the increasing number of publications that are emerging from around the globe linking KD to an infectious trigger. As a unique coincidence, Dr. Tomisaku Kawasaki, the legendary Japanese pediatrician after whom the disease is eponymously named, passed away recently [3].

Despite intense research efforts spanning over many decades, the exact etiology remains unknown. The etiological hypotheses include a KD-specific RNA virus, superantigen-mediated illness, and tropospheric winds transporting infectious or toxic agents. Environmental triggers, especially infections, are believed to trigger the disease in genetically susceptible individuals [4]. An important clinical observation that supports this hypothesis is the occurrence of KD in siblings who have disease onset in close temporal proximity to each other. Another clinical observation favoring this hypothesis is the older sibling often presenting with KD first—probably a reflection of greater likelihood of exposure to an infectious trigger. Despite an increased risk of developing KD, siblings of index cases may not be diagnosed due to atypical or incomplete presentations. Incomplete or “atypical” KD patients seem to be at a higher risk of developing coronary artery abnormalities (CAAs) which may be partly due to delays in diagnosis [5].

In this manuscript, we have reviewed the published literature on siblings with KD who developed the disease in close temporal proximity to each other. This may help in answering the long-standing question regarding the exact etiology of KD.

Search strategy

There have been only a few case reports and retrospective studies that have described the occurrence of KD in siblings within a short time period. Two authors (AZB, DB) independently performed a literature search using PubMed/Medline and Google scholar databases. The keywords included in the search were “Kawasaki disease” and “sibling,” “Kawasaki disease” and “twin,” and “Kawasaki disease” and “family.” The term “Mucocutaneous lymph node syndrome” was also used in place of “Kawasaki disease” for literature search and used in combination with sibling, twin, and family for performing a detailed literature search. All articles describing KD in siblings (irrespective of the time frame) were selected for full-text reading. English language articles describing the occurrence of KD in siblings within a month of each other were selected for our non-systematic review (Table 1). English language reports describing cases published in other languages were also included in the review. References of retrieved articles were also cross-checked. Reports where interval between the onset of KD in sibling pairs was not specified were excluded from this review. We were unable to retrieve the abstract and full text of one article reporting simultaneous KD in a twin pair (Table 1, S. No. 17) [22].

Table 1 Review of English language reports of KD in siblings
Full size table

Clinical presentation of Kawasaki disease

Kawasaki disease presents with a unique symptom complex of fever accompanied by (a) non-exudative conjunctival injection, (b) polymorphous rash, (c) unilateral cervical lymphadenopathy, (d) changes in lips or oral mucosa, and (e) extremity changes [23]. This characteristic symptom complex is included in the American Heart Association (AHA) clinical criteria for diagnosis of KD, with “complete” or “typical” KD patients fulfilling at least 4 of the abovementioned clinical features in addition to fever. Patients presenting with fever and only 2 or 3 classical manifestations are labeled as “incomplete” KD. Despite being so-called “incomplete” KD, coronary artery involvement may, in fact, be higher in these patients [24]. Unless detected early and treated promptly, the long-term consequences in patients with incomplete KD can be grave [25]. Incomplete KD was recognized even in the first published manuscript on KD [26].

Etiology of KD—a brief overview

Infections have come a long way as the probable cause eliciting an aberrant immune response leading to KD in genetically predisposed individuals. In-depth studies of autopsy specimens have implicated a specific RNA virus in causation of KD [27]. Several other infectious triggers that have been implicated include streptococci, staphylococci [28], Yersinia [29], and a number of viruses [30].

Epidemiological studies on KD have also revealed a striking association with the pattern of tropospheric winds, which have been hypothesized to transport Candida spp. and trigger KD through mechanisms that are not clearly understood [31,32,33]. In fact, Candida albicans water-soluble fraction has been shown to result in a KD-like illness in murine models [34].

Analysis of sibling pairs with KD is useful as it helps to reflect on these individual etiologies. The fact that the prevalence of KD is higher in some ethnic groups (e.g., Asians) led to the identification of genes attributed to increased susceptibility to the disease. Among these, single-nucleotide polymorphisms in CD40L, inositol 1, 4, 5-trisphosphate 3-kinase C (ITPKC), and several interleukin genes have been extensively studied [35]. Shimizu et al. also reported the association of transforming growth factor (TGF)-beta 2 (TGFB2), TGF-beta receptor 2 (TGFBR2), and SMAD3 polymorphisms with KD and the development of coronary artery lesions [36]. A linkage study by Onouchi et al. identified 10 chromosomal loci in siblings with KD which had positive linkage signals. Among these, 12q24 region had the most significant association [37]. However, these results need to be replicated from other geographical regions.

How common is KD in siblings?

The occurrence of KD in siblings was first reported by Tominaga et al. in 1977. They described 51 sibling cases, two-thirds of whom had an interval of < 1 week between onset of disease [38, 39]. The first English language report of KD presenting simultaneously in siblings appeared in 1978 [6]. Detailed epidemiological studies conducted in Japan and Korea usually describe a sibling incidence rate of KD of about 2% and 0.2% respectively [10, 11, 40,41,42]. Though these studies reported siblings with KD, the exact interval between disease onsets has not been clearly reported, and thus, these studies were not included in the present review.

Clinical features unique to KD in siblings

Irrespective of the time duration between disease onset, siblings appear to be at a significantly higher risk of developing KD [11]. Siblings may have incomplete KD or atypical KD [17] which may go unnoticed, and these children may go on to develop coronary artery abnormalities despite normal initial echocardiography [18, 19, 21]. There are several reports of incomplete KD in siblings (Table 1). Fever of any duration in siblings of KD, especially presenting within a short time interval after the index case, should alert the treating physician to a possibility of KD. These children should be actively screened for KD by a targeted history and detailed physical examination including subtle clinical pointers like perianal peeling, BCG site reactivation, and chromonychia. Other investigations that are helpful in these circumstances include assays of N-terminal pro-brain natriuretic peptide (NT-proBNP) assay and a detailed echocardiography. This would ensure timely treatment and decrease the risk of developing CAAs. In addition, these children need to be carefully followed up as CAAs may not be evident on initial echocardiography (Table 1). Data on coronary artery status were available for 13 sibling pairs only. Index patients developed CAAs in 3 pairs (3/13, 23%), and the sibling developed CAAs in 4 pairs (4/13, 31%). Additionally, only aortic root dilatation developed in the sibling in one pair (1/13, 8%).

KD in siblings—guide to etiology?

In addition to implications in disease management mentioned above, the study of KD in siblings provides a valuable opportunity to analyze the etiology of KD. This is because the siblings are likely to have similar genetic backgrounds and exposure to environmental risk factors. Analysis of tropospheric winds transporting infectious or toxic agents in relation to the occurrence of KD reveals a probable incubation period of 6–48 h, thereby reflecting a host response to the antigen or toxin rather than an infection per se [32]. In such circumstances, KD would probably occur simultaneously in siblings. Such instances have also been reported in the literature. It has been noted that the older sibling often presents first with KD and there is a time lag in disease onset in siblings with KD. This could be a result of the older sibling getting infected at school/daycare and passing on the putative infectious agent to the younger sibling at home. The time lag would then be explained by the incubation period seen in infectious diseases. Many patients with “infection triggered” KD have a proven infection. This holds true even for sibling pairs (Table 1). Some of these “infection triggered” KD cases may be mediated through superantigens or heat shock proteins rather than through the infectious agent directly [28]. Of more than 200 sibling pairs reviewed, a specific infectious trigger was documented in only 3 pairs which included streptococcus, adenovirus, and parvovirus. In one pair, throat swab showed growth of Haemophilus influenzae. However, its role in directly triggering the disease was conjectural. When siblings of a patient with KD develop a febrile illness in close temporal proximity, the attending pediatrician should be alert to the possibility of an infection that may also trigger KD [13].

Conclusion

KD needs to be suspected strongly in febrile siblings of children recently diagnosed to have KD. Studies on the etiopathogenesis of KD need to put a greater emphasis on sibling pairs who develop KD in close temporal proximity to each other. Detailed genome-wide association studies need to be conducted in such situations. Recently, KD has been reported in association with novel coronavirus (SARS-CoV-2/COVID-19) infection in children [43]. It may be prudent to look for signs of KD in siblings of patients with COVID-19 who develop a febrile illness.