Discussion:
The incidence of DKA amongst all pediatric age groups is rising, but especially in children less than five years of age [5]. In fact, DKA as the presentation of DM is much more common in children less than five years of age than any other age group, with up to 36% presenting with ketoacidosis [20, 21]. DKA management may necessitate CVC placement when obtaining peripheral access is difficult and infusion of medication and fluids is needed emergently. DKA is often associated with severe dehydration making it difficult to obtain IV access quickly, especially in children [22]. Adipose tissue distribution, smaller veins, and dehydration only enhances the difficulty [23].
Data from this study confirmed that children with DKA utilizing CVCs for disease management are at high risk for VTE development (~50% prevalence), and that children with DKA undergoing CVC placement are at a greater risk for developing CVC-associated VTE than those of a similar age group with CVC placed for any other reason. Our data also showed that children with DKA undergoing CVC placement are younger than those that do not (median age of 17 months vs. 33 months). Yet, assessment of both laboratory and demographic data amongst this patient population did not reveal any predictive indicators for VTE. In fact, the only significant difference between those with VTE and without VTE was length of hospital stay. This difference could be attributed to the acuity of the patients with VTE development. However, length of CVC use and degree of acidosis at presentation in the two patient groups was similar. A possible reason for the longer hospital stay was for VTE education and management.
At our institution from January 2002 through April 2007, there were 13 CVC placements for DKA management out of the 60 admitted patients. Specifically between February 2007 and April 2007, five of the seven patients admitted with DKA underwent CVC placement and three of these patients developed VTE. Only four of the 82 patients admitted for DKA underwent CVC placement after April 2007 and none had CVC-associated VTE. The decrease in CVC placement and CVC-associated VTE in this patient group is likely multifactorial. One patient had initial fluid resuscitation via an intraosseous tibial access at a referring hospital and had an external jugular CVC placed on arrival at our institution for continuing management. This patient’s VTE risk may have been lessened by fluid resuscitation prior to CVC placement. Also, fewer CVC were placed for acute DKA management likely due to recent institutional experiences and increased awareness of the VTE risk in children with DKA undergoing CVC placement [15, 18, 19]. Two of the five patients that presented after April 2007 received continuous unfractionated heparin for VTE prophylaxis. Given the lack of identifiable predictive VTE indicators and very high risk (>50% in our population), anticoagulation prophylaxis should be strongly considered for all young children with DKA undergoing CVC placement.
The pathophysiology leading to a prothrombotic state in patients with DKA is not clearly understood. Possible mechanisms include that DKA is an acute worsening of a chronic hyperglycemic state. Both acute and chronic hyperglycemia are pro-inflammatory states, which has a known association with a prothrombotic state [24]. Studies have shown that critically ill non-diabetic children with hyperglycemia are at higher risk for VTE development than critically ill children without hyperglycemia [25]. DKA has also been associated with increased cortisol levels and growth hormone levels [24]. Unprovoked VTE have been described in patients with hypercortisolism or elevated growth hormone levels, suggesting an associated prothrombotic state [26]. Plasminogen has also been shown to be glycated in diabetes leading to decreased plasmin generation, impaired protein activity, and, subsequently, hypofibrinolysis [27]. More dedicated research is necessary to fully understand the mechanisms involved in the prothrombotic state associated with DM and DKA.
Using a retrospective chart review for data collection does have limitations. Thrombophilia testing was performed at the discretion of each provider, and thus was not always performed. Radiographic investigation for CVC-associated VTE was limited to patients that had VTE symptoms. Nevertheless, these limitations apply to both the DKA group and the control groups. Determining length of CVC placement was difficult in the non-DKA group using a retrospective chart review. We focused on VTE development within three months of CVC placement in an effort to capture any CVC-associated thrombus.