- Top of page
Summary: Purpose: To analyze effects of different types of seizures and nonepileptic events as well as effects of seizure duration and lapse between the time of seizure and blood collection on serum prolactin level and peripheral white blood cell (WBC) count.
Methods: We prospectively collected blood samples from all patients admitted to our Epilepsy Monitoring Unit at baseline and after an event. Blood samples were analyzed, and serum prolactin level and WBC count were determined. Statistical analyses were performed to evaluate the relation of each type of seizure, its duration, and time lapse between a seizure and collection of blood sample to the serum prolactin level and peripheral WBC count.
Results: Serum prolactin level increases above twice the level at baseline after a complex partial seizure or a generalized seizure. Peripheral WBC count is elevated above the upper limit of normal in about one third of cases after a generalized seizure. In generalized seizures, the length of a seizure is positively associated, whereas the lapse time between the seizure onset and blood draw is negatively correlated with the increase in WBC count. Thus the longer the seizure and quicker the blood draw, the higher the WBC count.
Conclusions: We conclude that complex partial or generalized seizures are associated with an increase in serum prolactin level. Peripheral WBC count increases significantly after a generalized seizure and is probably transient in nature.
Serum prolactin level is a well-established test to help differentiate epileptic seizures from nonepileptic events (pseudoseizures). Elevation of serum prolactin level is common after complex partial and generalized tonic–clonic seizures and to a lesser extent after simple partial seizures (1,2). However, some have questioned the validity of serum prolactin levels in differentiating nonepileptic from epileptic events (3,4).
It is generally believed that status epilepticus and to some degree a generalized tonic–clonic seizure may induce leukocytosis (5). However, we did not find any prospective studies evaluating the relation of leukocyte count to a single seizure. The cause of this increase in the leukocyte count also is not well studied.
The white blood cell (WBC) count increases after rigorous exercise (6). We hypothesize that elevation of WBC count after a seizure is a result of muscular activity during the seizure. If that is the case, the elevation in WBC should be directly related to the length of generalized tonic–clonic activity. We studied the relation of leukocyte count to the type of seizure, the length of the seizure, and lapse of time between the seizure and collection of the blood sample.
- Top of page
The data were collected prospectively, but analyzed retrospectively. The Wayne State University Institutional Review Board for human investigations approved the study. The patients were admitted to the Epilepsy Monitoring Unit (EMU) at Harper University Hospital for continuous EEG and CCTV monitoring. Each event was classified according to criteria for seizure type published by the International League Against Epilepsy. A serum prolactin level and complete blood count (CBC) were obtained at baseline and immediately after a seizure. The baseline studies were obtained on admission to the EMU, usually several hours after admission, as it took ∼1.5–2 h to apply the scalp electrodes. Any seizure during that time was reported. When the WBC count exceeded the upper limit of the normal range, or serum prolactin level increased more than twice the baseline level, they were considered abnormal.
The length of each seizure was determined by reviewing both the video and electrographic data of the event. We also recorded lapse time between the event and collection of blood to evaluate its effect on the WBC count. Seizures lasting <10 s, or where there was a delay >120 min between the seizure and blood collection, were not considered. Brief generalized seizures were excluded from the study to avoid inclusion of brief myoclonic seizures that were classified as generalized seizures. Only events for which baseline and postevent WBC and prolactin levels were available were included in the study.
- Top of page
We studied a total of 340 events in 89 patients. One hundred two events were classified as nonepileptic events (NES), 38 as simple partial seizures (SPS), 109 as complex partial seizures (CPS), and 91 as generalized tonic–clonic seizures or complex partial seizures with secondary generalization (GEN). A serum prolactin level was obtained both at baseline and after a seizure for 189 events. A WBC was obtained both at baseline and after a seizure for 193 events. Seizure classification and baseline plus both postevent WBC and prolactin levels were available for 174 events.
The average lapse times for drawing blood for various events shown in Table 1. The differences between average Lapse times for the four groups were not statistically significant [F(3, 219) = 1.21, p = 0.31, NS]. Overall, 90% of subjects had blood drawn within 40 min. Fewer than 1% of subjects had Lapse time >74 min, and longest Lapse time was 93 min.
Table 1. Lapse time for collection of blood for various events
|Lapse time (min)||Seizure type|
Table 2 summarizes the data regarding WBC and prolactin level, and also indicates whether WBC was abnormal (>10.5 K/mm3), or prolactin was abnormally increased (>2× from baseline), after the seizure. A χ2 test of independence indicated a significant relation between seizure type and elevated WBC [χ2 (3, N = 174) = 37.65, p < 0.01]. Inspection of Table 1 shows that WBC was increased >10.5 in 36% of GEN seizures, and 7% of CPSs. NESs and SPSs showed no abnormal increase in WBC. A χ2 test of independence indicated a significant relation between seizure type and abnormally elevated prolactin [χ2 (3, N = 174) = 31.80, p < 0.01]. Inspection of Table 2 showed that 47% of GEN seizures, 34% of CPSs, and 11% of SPSs showed abnormally elevated prolactin. Only 2% of NES events showed this increase.
Table 2. Frequencies (precentages) of increased WBC count and serum prolactin level by seizure type
| ||Seizure type|
|WBC|| || || || || |
| Increaseda||0 (0)||0 (0)||4 (7)||13 (36)||17|
| Normal||55 (100)||27 (100)||52 (93)||23 (64)||157|
|Prolactin|| || || || || |
| Increasedb||1 (2)||3 (11)||19 (34)||17 (47)||40|
| Normal||54 (98)||24 (89)||37 (66)||19 (53)||134|
If one considers absolute value of the serum prolactin level, the results are different (see Table 3). The majority of the patients have serum prolactin level above the upper limit of normal (>18.1 ng/ml) after a SPS, CPS, or CPS with secondarily generalization. The actual percentages are 55% for SPSs, 66% for CPSs, and 86% for GEN. However, as shown in Table 3, patients with NESs (14%) also show an abnormally high value of serum prolactin after an event. At baseline, serum prolactin level was high in 5% of NESs, 10% of SPSs, 33% of CPSs, and 16% of GEN subjects.
Table 3. High serum prolactin level and its relation to various types of seizures
| ||Seizure type|
|Baseline serum prolactin|| || || |
| Increased (>18.1 ng/ml)a||4 (5%)||3 (10%)||26 (33%)||12 (16%)|
| Normal (≤18.1 ng/ml)||72 (95%)||25 (90%)||52 (67%)||64 (84%)|
|Postictal serum prolactin|| || || |
| Increased (>18.1 ng/ml)||11 (14%)||11 (55%)||42 (66%)||67 (86%)|
| Normal (≤18.1 ng/ml)||70 (86%)||9 (45%)||22 (34%)||11 (14%)|
There was a significant relation between seizure type and seizure duration (SZDUR; F(3, 260) = 10.40, p < 0.01). Post hoc Tukey tests showed that NESs had a much longer mean seizure duration (318 s) than did SPSs, CPSs, and GEN types (63, 98, and 116 s, respectively), which did not differ. There was no relation between seizure type and LAPSE (time between the seizure and blood collection).
Mean change in WBC (postevent WBC minus baseline WBC; DWBC) for the seizure types is shown in Fig. 1. Overall, there was a significant relation between seizure type and DWBC [F(3, 189) = 27.81, p < 0.01). Post hoc Tukey tests showed that the GEN seizures showed a mean increase in WBC that was significantly greater than that of any other seizure type. CPSs also showed an (smaller) increase, but it was significantly greater than only that shown by NESs.
Mean change in serum prolactin (postevent prolactin minus baseline prolactin; DPRL) for the seizure types is shown in Fig. 2. There was a significant relation between seizure type and DPRL [F(3, 185) = 14.83, p < 0.01). Tukey tests showed no differences between GEN, CPS, and SPS, but all three seizure types showed a significant increase in prolactin relative to NES.
Finally, the linear relation between DWBC and DPRL with SZDUR and LAPSE (predictors) were evaluated by multiple regressions within each seizure type. These analyses showed that for GEN seizures, DWBC was positively related to SZDUR [t(43) = 2.36, p < 0.05] and negatively related to LAPSE [t(43) = −3.41, p < 0.01]. For GEN events, increased WBC was associated with longer seizures and shorter intervals between the seizure and blood collection. DWBC was not related to SZDUR or LAPSE in any other seizure type. DPRL was not related to SZDUR or LAPSE in GEN, CPS, or SPS types. It was positively related to SZDUR for NES [t(30) = 2.79, p < 0.01], although, as indicated earlier, 98% of the events did not result in abnormal increase of prolactin level.
- Top of page
Our study replicates previously reported findings that serum prolactin level increase after a seizure, but does not increase after a nonepileptic event (1,2). However, it is important to emphasize that the increase of serum prolactin level is significant only when it is greater than twice the baseline level. An isolated postictal prolactin level may be misleading, as ≥14% of nonepileptic events may be associated with a serum prolactin level that is higher than the upper limit of normal. Our results also indicate that this increase in serum prolactin level after a seizure may be less frequent than previously reported (7,8).
The main focus of our study was to analyze the association of peripheral WBC count and seizures. In reviewing the literature, we did not find any systematic analysis of WBC counts in relation to a single seizure. Our analysis shows that slightly more than one third of generalized seizures are associated with a significant increase in WBC count. Complex partial seizures were associated with an increase in WBC count very infrequently (7%), whereas none of the events classified as NES or SPS was associated with high WBC count.
The seizure duration was not associated with elevation of WBC count in partial seizures. There is a clear correlation between the length of a seizure and increase in WBC count in generalized seizures, controlling for lapse time. This indicates that the longer the generalized seizure, the higher the WBC. The lapse time between the seizure onset and blood draw was negatively correlated with WBC count. This suggests that the increase in the WBC count is related to the seizure and is transient. That the mean change in WBC count is related to lapse time only in the generalized group is not surprising. This group shows a greater elevation of WBC than all others (CPS also showed an elevation, but it was only greater than the NES group), and one needs an increase in WBC for there to be a decline over time (i.e., significant negative relationship with lapse time).
The exact mechanism for the increase in WBC count after a seizure is not known. One possibility is demargination of leukocytes secondary to muscular activity during a generalized seizure. Other possibilities include release of catecholamine or cortisol during a seizure (6,9,10).
Our findings are consistent with those patients with status epilepticus, in whom a prolonged seizure or repeated seizures are likely to cause excessive muscular activity and leukocytosis (5). Toyosawa (9) demonstrated in an animal model of electrically induced convulsions that the increase in WBC count after a convulsion is bimodal. The first peak occurs within 20–30 min of the seizure, and a second peak occurs at 4 h. Blocking the muscular activity by administrating a muscle-paralyzing agent before seizure induction prevents the immediate increase in WBC count. The second peak is not altered by the use of a muscle relaxant before the seizures. However, pretreatment with 6-OHDA (6-hydroxydopamine), resulting in selective destruction of adrenergic nerve terminals, caused an abolition of the second peak. Thus the second peak of WBC increase appears to be hormonally mediated, perhaps through the adrenergic system. In another study, electroconvulsions in rabbits caused a polymorphonuclear leukocytosis. Daily administration of electroconvulsions did not result in tolerance to this response, but prior administration of propranolol blocked the leukocytosis, suggesting the adrenergic system's involvement in mediation of this response (10).
The relation between muscular activity and increase in WBC count has been studied in athletes. In human exercise physiology data, propranolol administration before exercise did not consistently alter the increase in WBC count. The peak in the WBC count occurring at 3–4 h is thought to be related to cortisol-mediated leukocytosis (6).
A single generalized tonic–clonic seizure is known to cause an increase in circulating epinephrine and norepinephrine, and leukocytosis can be considered to be due to the demargination effects of these circulating catecholamines (5). However, partial seizures, especially CPSs, can be considered to be emotionally stressful events, and CPSs also can result in direct autonomic stimulation. In that case, CPSs should be associated with release of catecholamines and, in turn, should increase the leukocyte count This may be or may not be true, as effects of muscular activity independent of circulating catecholamines have not been assessed. In addition, the demargination effect will result in leukocytosis without bandemia, although this was not systematically studied in our study, as differential WBC count was not obtained at baseline and after each event. At times muscular activity during a nonepileptic event may be significant, and it may mimic generalized tonic–clonic seizure. In spite of that, we did not find a single instance of increased WBC count after an NES. However, we did not quantify muscular activity during the event, and most of the events classified as NESs were not associated with violent muscular activity.
Serum cortisol level increases after a generalized or partial seizure or a pseudoseizure (11,12). Our findings, that the WBC count increases after a generalized seizure and to a lesser extent CPS but does not increase after an SPS or NES, suggest that the increase is not due to cortisol release. However, if the cortisol-mediated increase in WBC may occur 3-4 h after the event, we might not have detected it, as we did not collect blood samples then.
A further study with measurement of serial WBC counts as well as serum cortisol levels and catecholamine levels may help to define the underlying mechanisms of seizure-related increase in WBC count. Studies have shown robust increases in serum prolactin levels after electroconvulsive therapy (ECT) (13). Similarly, studies of the relation of peripheral WBC count and ECT also may provide insight into the underlying mechanisms of postictal leukocytosis. Because muscles are paralyzed during ECT, the demargination effect of muscular activity on the WBC count would be eliminated.