Table 3 Overview of the USFD parameters related to outcome

 Age (Core)

14

Age is described as mean, median or range: Mean age 3 years and 5 months-47.5 years [3,4,5,6,7,8, 12, 16]; median age 2–2.2 years [9, 10]; Range 1–60 years [11, 13,14,15].

One publication shows that survivors are significantly younger than non-survivors (38.9 ± 12.6 versus 56.6 ± 18.7; p = 0.03) [4].

 Date and time of day of incident (Core)

6/9

Six publications describe the time of day [3, 5, 6, 8, 13, 14] and nine publications describe the season the drowning occurred in [3, 5,6,7,8, 10, 11, 13, 14]. Four publications describe that 39–100% of drownings occur during the daytime [3, 5, 8, 13] One publication reports that 67% of drownings occur between 12 AM and 8 PM and 33% of drownings from 9 PM to 11 AM [6]. Forty-seven to One hundred percent of drowings are reported to occur between springtime and the end of the summer [3, 5,6,7,8, 10, 11, 13].

One publication shows that outcome after drowning is significantly better in winter compared to other seasons: odds ratio 4.6 (1.4–15.1), p = 0.013 [10].

 Witnessed (Core)

10

The drowning event is not witnessed in the majority of cases (58–81%) in four publications [3, 4, 8, 9]. In another five publications the drowning is witnessed in the majority [5,6,7, 10, 16]. In one publication the drowning was witnessed in 100% (by fellow victims) [14].

One publication shows that survivors are more likely to have had a witnessed drowning event than non-survivors (76 versus 61%; p = 0.036) [5].

 Approximate water temperature (Supplemental)

6

In one publication a median water temperature for survivors of 20.4 (Interquartile range 13.7, 27.0) and 20.0⁰ C (interquartile range 8.6, 23.8) was reported (p = 0.184) [8]. In three publications a large range (0–28⁰ C) was reported [3, 10, 16]. In one study the water temperature was 2⁰ C for all victims [14]. In one publication the water temperature was described as warm (14%), cold (84%, or ice-cold (3%) [6].

In one publication the water temperature was lower for 24 h survivors among victims treated with ECLS (p = 0.04) [16].

 Time of submersion (Supplemental)

10

The duration of submersion is described in different ways (median, mean, ordinal) but ranges from < 1 to 45 min [3, 5,6,7,8,9,10,11,12]. One publication described a submersion time of 10 min or more in 2% of the victims [15].

A longer duration of submersion is significantly associated with bad outcome in four publications [5, 8,9,10]. Two publications report no significant difference [3, 7].

 Resuscitation before EMS arrived (Core)

10

Resuscitation attempts before the arrival of EMS varies between 24 and 93% [3,4,5,6, 8,9,10,11]. In one publication this parameter was described in 19% of the cases, of which 79% were resuscitated before EMS arrival [15]. In one publication BLS was immediately started after removal from the water in 100% of the victims by police officers or firefighters before ALS providers arrived [16].

One publication shows that drowning victims that survive have significantly more bystander CPR than non-survivors (57 versus 17%; p = 0.03) [4]. A significant relation between bystander resuscitation and outcome is not found in three publications [5, 8, 10].

 Oxygen saturation, temperature, blood pressure, pupillary reaction (Supplemental)

6

In one publication at least one of these four parameters information is available in 4 out of 343 victims [6]. Hypothermia is reported in four publications [3, 12, 14, 16]. Pupillary reactions are described in three publications [3, 9, 12].

In one publication unresponsive and dilated pupils is significantly related with bad outcome (p < 0.001) [9]. In one publication the first prehospital core temperature was lower in the 24 h survivors among patients treated with ECLS (p = 0.07) [16]. In this same publication the association between a first prehospital core temperature of ≤26⁰ C and serum potassium level between 4.2 and 6.0 identified 24 h survivors among patients treated with ECLS with 100% sensitivity (95% CI: 28–100%) and specificity (95% CI: 71–100%).

 Time of first EMS resuscitation attempt (Core)

3

The mean time interval is described in two publications [5, 7].

In another publication this parameter is included, but the information is not available in any victims [6].

The time of first EMS resuscitation is found to be significantly associated with outcome in one publication: 11.2 ± 5.6 min in survivors versus 21.4 ± 12.8 min in non-survivors; p = < 0.001 [5].

 Vital signs (Core)

8

Vital signs are absent in 20 to 100% of the victims on arrival at the ED [3,4,5, 7, 9, 10, 14]. In one publication all victims had asystole at arrival at the ED [16].

In one publication resuscitation at arrival in the ED is negatively associated with outcome (p < 0.001, OR 0.03, 95% CI 0.01–0.13) [9]. In the same publication hypothermia is significantly related to bad outcome (p < 0.001, OR 18.00, 95% CI 3.35–96.74). In one publication 24 h survivors among patients treated with ECLS had a significantly lower in hospital initial core temperature (p = 0.004) [16].

 Arterial blood gas analysis, if unconscious or SaO2 < 95% on room air (Core)

8

Hypoxemia, acidosis, and hypercarbia are common findings [3, 5, 7,8,9,10, 14, 16].

The more severe the acidosis, the worse outcome is (p < 0.001–0.014) [8,9,10]. In one publication an initially lower pH relates to hospital mortality (p = 0.008) [7]. Drowning cardiac arrest victims have a higher initial pCO2 compared to non-drowning victims in cardiac arrest (p < 0.001). Endtidal CO2 after 1 min of CPR (p 0.02) and the final endtidal CO2 (p < 0.001) were independent factors for survival [10]. Less negative base excess is related to better outcome. (p < 0.001–0.001) [8, 10].

 Initial neurological status (Core)

8

The GCS is 3 at the ED in all patients in three publications (in only one patient in one of these studies a palpable pulse was reported) [3, 5, 7]. In one publication the median Glasgow Coma Score (GCS) ranged between 11.5 in survivors and 3 in non-survivors [8]. In one publication the GCS was described as < 5 (41% good outcome, 18% neurologic sequelae, 41% death) or ≥ 5 (98% good outcome, 2% death) [9]. In one publication the median GCS was 3 [10]. In one publication the GCS ranged from 3 to 15 [14]. One study described that none of the victims had clinical signs of life [16].

A low GCS is significantly associated with bad outcome in three publications (p < 0.001) [8,9,10].

 Pupillary reaction (Supplemental)

4

In three publications, fixed and dilated pupils were reported in 47%, 95% and 100% of victims [5, 7, 14]. In one publication pupillary reactions are described as reactive (n = 44), sluggish (n = 6), unreactive not dilated (n = 6), or unreactive dilated (n = 5) [9].

Unreactive dilated pupils in the ED are significantly related to bad outcome in one publication (p < 0.001; OR 0.01; 95% CI 0.04–0.23) [9].

 Airway and ventilation requirements (Core)

7

In six publications 100% of the patients are ventilated mechanically or manually [3, 5, 7, 9, 10, 16]. In one publication it was reported that one patient was intubated and ventilated [14].

Intubation at the ED (p = 0.002) is significantly related to bad outcome in one publication [9].

 Serial neurological function (admission, 6 h, 24 h, 72 h, discharge) (Supplemental)

3

In one publication myoclonic or seizure activity (including treatment with medications), loss of pupillary response, absent motor response to pain, somatosensory evoked potentials (SSEPs), and the use of brain imaging are described [7]. One publication only describes the use of a CT scan [12]. One publication describes the use of electroencephalographic recordings, SSEPs, magnetic resonance imaging and the use of biomarkers (neuron specific enolase and protein S100B) [14].

Neurological function testing, somatosensory evoked potentials (SSEPs), brain imaging (computed tomography or diffusion-weighted imaging) and neurological examination of motor response to motor response to pain after 3 days, are significantly related to bad outcome in one publication [7].