Urinary Neutrophil Gelatinase-Associated Lipocalin (NGAL) and proteinuria predict severity of acute kidney injury in Puumala virus infection

In this study, we provide first evidence that uNGAL, as a point of care variable determined in emergency room (ER) patients having symptoms compatible with PUUV, is a good predictor of the severity of AKI due to PUUV infection. Although markers of proteinuria (uACR, uPCR and semiquantitative proteinuria determined by dipstick test) also correlated with the severity of AKI, these associations were weaker and uNGAL preserved a considerably stronger association with peak creatinine in models adjusted for uACR or uPCR. The highest discriminatory power for severe AKI, defined as AKIN stage 3, among the investigated variables was observed for uNGAL.

Our study delivers a proof-of-principle that predicition of disease severity is feasible in NE. Most patients with NE present with unspecific flu-like symptoms such as headache and fever, but often develop severe flank pain and oliguria soon afterwards. Laboratory analysis usually demonstrates thrombocytopenia and elevated plasma creatinine concentration at this stage. However, creatinine is a tardy and unreliable marker of acute kidney injury. Striking differences in peak creatinine concentration and associated problems of AKI such as electrolyte disturbances and uremic symptoms are observed in patients. Weight change during disease may indicate fluid retention and could therefore be an additional marker of disease severity reflecting capillary leakage [13, 14].

In this study, we show that uNGAL, and to a lesser degree markers of proteinuria, are able to predict severity of AKI. Severe AKI in turn can be associated with complications and the need for intensive monitoring of fluid- and electrolyte status. In the case of uNGAL, its associations with length of hospital stay and weight change during hospital stay further underlines its role in predicting disease severity. Thus, determination of uNGAL in patients with suspected NE right upon first presentation could guide the clinician’s decision about hospital admission. Our data suggest that an uNGAL >130 ng/ml (positive predictive value 90 %, cf. Table 5) is associated with severe AKI prompting hospital admission for further monitoring. In contrast, patients with low NGAL (<30 ng/ml) are at a relatively low risk for severe AKI, and their care could be continued in an outpatient setting. Adjustment for urinary creatinine did not improve the correlation, as also shown in a previous work [10].

Whereas thrombocytopenia also associated with length of hospital stay in our study, it was not correlated with other variables of disease severity. Since thrombocytopenia is typical in early phases of NE, this association is probably confounded by the time of presentation in the ER relative to disease phase. Patients who tended to present at an earlier disease phase would have lower platelet counts and would therefore have to be observed longer until resolution of AKI. Thrombocytopenia, however, could prove to be a better predictor of disease severity if determined early in the disease course, as shown by Rasche et al. [15].

The underlying pathomechanisms of AKI in PUUV infection is not fully understood. Although the virus primarily replicates in endothelial cells, it does not cause major cytotoxicity or cell death [2]. Rather, it seems to transiently disrupt the integrity of cell-to-cell junctions in renal podocytes, tubular epithelial and glomerular endothelial cells [16]. However, the contrast of clear interstitial edema and occasional tubular damage with relatively spared glomeruli seen in renal biopsy specimens suggests that the majority of renal pathology is due to tubulointerstitial injury [17–19]. NGAL predominantly accumulates in epithelial cells of proximal tubules where it is thought to play a role in induction of reepithelialization [20, 21]. Therefore, the observed strong association of uNGAL with peak plasma creatinine concentration, which was independent of proteinuria, supports the predominantly tubulointerstitial injury in NE in two ways. NGAL produced by neutrophil cells is freely filtrated and reabsorbed at proximal tubular sites [22]. Hence, proximal tubular injury would result in increased NGAL excretion by reduced tubular reabsorption. Furthermore, NGAL mRNA expression and production is upregulated in the distal convoluted tubule and the collecting duct during tubular stress [23]. These mutually enhancing effects and the marked tubulointerstitial injury in NE [17] could explain the strong association of uNGAL with the severity of AKI.

Several other biomarkers have previously been associated with the severity of NE. These comprise plasma markers such as interleukin-6 [24], indoleamin 2,3-dioxygenase [25], pentraxin-3 [13], cell-free DNA [26] and the urinary biomarker GATA-3 [27]. One study found that the maximum leukocyte count, which has been seen 3–6 days after disease onset, correlated with fold-change in plasma creatinine [27]. As to our knowledge, association of NGAL and simple urine biomarkers such as uACR or uPCR with outcome variables has not been investigated before. uNGAL determined in the ER upon first presentation of the patient seems to be the clinically most relevant and robust predictor, which seems to be independent from disease phase and time of onset of symptoms.

The single most important limitation of our study is its retrospective aspect. In a university hospital setting, we also note that the patients in our cohort were admitted at a relatively late disease phase (median 6 days after the onset of first symptoms), which potentially explains why established early markers of disease severity such as thrombocytopenia did not perform well.

Nevertheless, our robust findings support the notion that uNGAL allows for prediction of severity of AKI, enabling a risk stratification of these patients upon first presentation in the ER for suspected NE. Future trials have to prospectively validate these results.