Seasonal variation in the clinical recovery of patients with Guillain Barré syndrome requiring mechanical ventilation

Transcription

1 Original Article Seasonal variation in the clinical recovery of patients with Guillain Barré syndrome requiring mechanical ventilation Kamath Sriganesh, Archana Netto 1, Girish B. Kulkarni 2, Arun B. Taly 2, Ganne S. Umamaheswara Rao Departments of Neuroanesthesia, 2 Neurology, National Institute of Mental Health and Neurosciences, 1 Department of Neurology, Bangalore Medical College and Research Institute, Bangalore, Karnataka, India Abstract Address for correspondence: Dr. G. S. Umamaheswara Rao, Department of Neuroanaesthesia, National Institute of Mental Health and Neurosciences, Bangalore , India. E mail: gsuma123@yahoo.com Received : Review completed : Accepted : Background: Guillain Barré Syndrome (GBS) has a variable clinical course. The influence of season on the rate of recovery has not been evaluated previously, despite documentation of seasonal variation in the occurrence of GBS. This study evaluated the influence of season on the rate of recovery from GBS. Materials and Methods: Records of 184 patients with GBS over a 10 year period were reviewed. Patients were divided into four groups depending on the date of admission: Q1 (March-May), Q2 (June-August), Q3 (September- November), and Q4 (December-February). Demographic characteristics and recovery characteristics (duration of mechanical ventilation, ICU and hospital stay, and time for recovery from the time of initiation of definitive therapy) were compared across the four quarters. Results: There was no significant difference in age, antecedent illnesses, treatment received, electrophysiological findings, and muscle power at admission across the four groups. Significant differences among various seasons were found with respect to duration of mechanical ventilation (23 ± 20, 36 ± 34, 27 ± 22, and 38 ± 28 days for Q1-Q4, respectively; P = 0.05), ICU stay (27 ± 22, 40 ± 37, 31 ± 23, and 43 ± 30 days for Q1-Q4, respectively; P = 0.05), hospital stay (42 ± 28, 55 ± 44, 47 ± 34, and 72 ± 54 days for Q1-Q4, respectively; P = 0.02), and time for recovery from treatment (15 ± 14, 29 ± 34, 18 ± 14, and 29 ± 20 days for Q1-Q4, respectively; P = 0.02). Conclusions: This study demonstrates a seasonal variation in the recovery of patients with GBS requiring mechanical ventilation. Patients admitted in Q1 have the fastest recovery and those in Q4 have the slowest recovery. Key words: Demyelination, Guillain Barré syndrome, mechanical ventilation, prodromal illness, recovery, seasonal variation Introduction Guillain Barré Syndrome (GBS) in general has a good prognosis. However, the clinical course may be quite Quick Response Code: Access this article online Neurology India Jul-Aug 2013 Vol 61 Issue 4 Website: PMID: *** DOI: / variable in different patients. Factors like rapidity of progression and certain specific electrophysiological characteristics are correlated with the outcome. [1,2] In a recent retrospective study, patients requiring mechanical ventilation and age were the significant predictors of Hughes scale score at discharge. [3] Despite advances in specific and supportive care, about a fifth of the patients have poor outcome (death or permanent disability). [4] In a large series involving 297 GBS patients, clinical recovery was seen in 82% of patients at 24 months and 16% had residual functional disability. [1] Recovery has been documented to be slower in patients with primary axonopathy compared to patients with demyelination or 349

2 mixed neuropathy. [1] Increasing age, preceding illness, and GBS disability score at 2 weeks after admission are the other factors correlated with inability to walk independently at 6 months. [5] Over several decades, we observed that the duration of intensive care unit (ICU) stay of patients with GBS requiring mechanical ventilation is related to the season of the year during which they get admitted. In this study, we aimed to validate this observation. Materials and Methods Case records of 184 patients admitted to neuromedical ICU between July 1997 and June 2007 with a diagnosis of GBS were reviewed. The study was approved by the Institutional Ethics Committee. Diagnosis of GBS was established by the National Institute of Neurological and Communicative Disorders and Stroke (NINCDS) clinical criteria for GBS and confirmed by electrophysiological criteria wherever possible (121/155 analyzed patients). All patients were managed by routine critical care comprising cardiorespiratory optimization including mechanical ventilation, monitoring for and management of infections, nutritional care, and physiotherapy. The decision to tracheal intubation and initiate mechanical ventilation was based on the clinical signs of respiratory distress, corroborated with spirometry and blood gas analysis in the majority of cases. Attempts to wean the patient off mechanical ventilation were made when the forced vital capacity (FVC) improved to around 15 ml/kg body weight. All patients were ventilated in a pressure controlled ventilation mode (PCV) to begin with. Weaning was accomplished through pressure support ventilation (PSV) followed by T piece trials. The data collected included: Age, sex, antecedent illness, duration of symptoms before ICU admission, day of tracheal intubation, day of commencement of mechanical ventilation, muscle power graded by the British Medical Research Council (MRC) scale, respiratory and bulbar function at the time of admission to ICU, interval between onset of illness and initiation of definitive treatment, details of ICU complications, and the findings of cerebrospinal fluid (CSF) and electroneuromyographic analysis. Information was also obtained on the definitive therapies instituted, such as plasmapheresis and intravenous immunoglobulin (IVIG). Variables linked to the rate of recovery such as duration of mechanical ventilation, ICU stay and hospital stay, and days for recovery were also noted. Definitions a. Days for recovery: From the day of starting definitive treatment to the day of starting the weaning process b. Duration of mechanical ventilation: From the day of starting mechanical ventilation to the day of total disconnection from the ventilator c. Duration of ICU stay: From the day of entry to the ICU to the day of exit from the ICU d. Duration of hospital stay: From the day of presentation to the hospital to the day of discharge from the hospital. Of the 184 patients, 18 patients who died in the ICU and 11 patients who were admitted with suspicion of impending respiratory failure, but did not need tracheal intubation/mechanical ventilation were excluded from the analysis. These patients were excluded as it was only aimed to study the cohort of patients who required mechanical ventilation. Mortality in GBS is generally due to pneumonia, dysautonomia, and elderly age. [6] The non survivors are likely to have died before they had enough time for maximal neurological improvement. Inclusion of such patients would have falsely lowered the duration of ICU stay and mechanical ventilation. The remaining 155 patients were included in the analysis. Patients were divided into four groups based on the quarter of the year in which they were admitted to the hospital: Q1 (March to May), Q2 (June to August), Q3 (September to November), and Q4 (December to February). The rationale for such division is as follows: Temperatures are highest during March-May in Bangalore, where this study was conducted. Summer peaks in the month of April. Monsoon starts in June and temperatures progressively decline from June to August. Lowest temperatures are recorded during December to February. Statistical methods Recovery characteristics of the four groups were compared by one way analysis of variance (ANOVA) with post hoc testing by Least Significant Difference (LSD) method. Comparisons of muscle power among the four groups were made by using Chi square test. Demographic and baseline parameters influencing the outcome were analyzed using Chi square test for non parametric variables and one way ANOVA for parametric variables. A P < 0.05 was considered significant. Statistical analysis was carried out using SPSS version 17 software. Results The data of 155 patients who required mechanical ventilation and eventually could be weaned off the ventilator were analyzed. There were 104 male and 51 female patients. Their mean age was 32 ± 21 years. The demographic characteristics of the patients and the details of therapy are shown in Table 1. Age and sex distribution of the patients were similar in all the four groups. All the patients received either plasmapheresis or IVIG. A few 350 Neurology India Jul-Aug 2013 Vol 61 Issue 4

3 r patients received both plasmapheresis and IVIG (n = 10); technical failure to obtain adequate plasma volume for pheresis was the common cause for the dual treatment in these patients. The other baseline factors that could have an influence on the neurological outcome including electrophysiological study were not significantly different across the four groups [Table 2]. The time spent by patients in the hospital prior to initiation of ventilation was similar in all the groups [Table 2]. Proximal and distal muscle power in both upper and lower limbs at admission to ICU was also similar across the groups [Table 3]. The distribution of ICU factors that have the potential to influence duration of ventilation and ICU stay are shown in Table 4. Table 5 shows the duration of mechanical ventilation, duration of ICU and hospital stay, and time for recovery in all the groups. Group Q1 had the shortest duration of mechanical ventilation, ICU and hospital stay, and time taken for recovery from treatment and time to walk (Q1 vs. Q4) as compared to Q2 and Q4. Mortality across groups in the total sample (not involving survivors alone) was not different. Discussion The present study demonstrated that in GBS, the time to recovery, duration of mechanical ventilation, duration of ICU stay, and hospital stay of patients show a seasonal variation. Patients admitted between March and May had the fastest recovery and those admitted between December and February had the slowest recovery. There are no previous studies that examined the seasonal variations in the recovery from GBS. In Table 1: Demographic characteristics (n=155) Variable Q1 [n=52] Q2 [n=40] Q3 [n=35] Q4 [n=28] Significance P values for post hoc analysis Age (y) 35±22 29±19 30±18 29±20 P=0.4 Q1 vs Q2=0.1 Q1 vs Q4=0.2 Gender (M: F) 30:22 26:14 29:6 19:9 df=3; χ 2 =6; P=0.1 Q1 vs Q2=0.5 Q1 vs Q4=0.4 Plasmapheresis (n=116) df=3; χ 2 =3.3; P=0.3 Q1 vs Q2=0.2 Q1 vs Q4=0.9 IVIG (n=45) df=3; χ 2 =4.6; P=0.1 Q1 vs Q2=0.3 Q1 vs Q4=0.4 Q1 March May, Q2 - June Aug, Q3 Sept Nov, Q4 Dec Feb Table 2: Baseline parameters that have a potential to influence the outcome Parameter Q1 [n=52] Q2 [n=40] Q3 [n=35] Q4 [n=28] Significance Post hoc analysis Antecedent event {AE} total; present (n) df=3; χ 2 =1.5; P=0.7 Q1 vs Q2=0.5 Q1 vs Q4=0.9 AE fever; present (n) df=3; χ 2 =1.3; P=0.7 Q1 vs Q2=0.7 Q1 vs Q4=0.4 AE respiratory infection; present (n) df=3; χ 2 =1.9; P=0.6 Q1 vs Q2=0.8 Q1 vs Q4=0.2 AE gastrointestinal infection; present (n) df=3; χ 2 =2.3; P=0.5 Q1 vs Q2=0.1 Q1 vs Q4=0.6 AE injection/vaccine; present (n) df=3; χ 2 =1.7; P=0.6 Q1 vs Q2=0.4 Q1 vs Q4=0.4 Duration of Antecedent event to onset of illness (days) 5.1± ± ± ±4.5 P=0.35 Q1 vs Q2=0.4 Q1 vs Q4=0.9 Bulbar palsy (n) df=3; χ 2 =1.3; P=0.7 Q1 vs Q2=0.9 Q1 vs Q4=0.3 SBC at admission (sec) 17±11 14±10 10±6 11± Q1 vs Q2=0.3 Q1 vs Q4=0.06 Respiratory rate at admission (bpm) 22±7 22±6 24±7 19± Q1 vs Q2=0.9 Q1 vs Q4=0.2 CSF cell count (cells.mm 3 ) 5±23 3±5 2±2 1± Q1 vs Q2=0.5 Q1 vs Q4=0.3 CSF protein (mg.dl 1 ) 132±115 93±79 128± ± Q1 vs Q2=0.4 Q1 vs Q4=0.4 Onset of illness to ventilation time (days) 7.6± ± ± ± Q1 vs Q2=0.5 Q1 vs Q4=0.3 Days in hospital before ventilation (days) 2.9± ± ± ± Q1 vs Q2 = 0.8 Q1 vs Q4 = 0.7 Onset of illness to treatment (days) 8.8± ± ± ± Q1 vs Q2=0.8 Q1 vs Q4=0.7 NCS Axonopathy (n) df=3; χ 2 =2.9; P=0.4 Q1 vs Q2=0.1 Q1 vs Q4=1 NCS Unexcitable (n) df=3; χ 2 =0.88; P=0.8 Q1 vs Q2=0.5 Q1 vs Q4=0.9 Q1 March May, Q2 June Aug, Q3 Sept Nov, Q4 Dec Feb, SBC Single breath count, CSF Cerebrospinal fluid, NCS Nerve conduction study, DOI Day of illness Neurology India Jul-Aug 2013 Vol 61 Issue 4 351

4 this study, we chose to use the duration of mechanical ventilation, ICU and hospital stay, and time taken for recovery as measures of rapidity of recovery from the illness. From a practical perspective, this association would provide information on the possible morbidity and duration of ICU and hospital stay during a given season of the year. Earlier studies have reported only seasonal variations in the incidence of GBS and not the rate of recovery. In a recent study involving 25 patients, Zaheer et al. had reported a bimodal incidence of GBS during April- May (24%) and July-August (32%) as compared to the other months of the year. [7] Similar increased incidence in summer has been reported by others. [8,9] An Italian report, however, did not find such seasonal variation Table 3: MRC muscle power grade at admission Season Muscle power [Median (Interquartile range)] Lower limbs Upper limbs Proximal Distal Proximal Distal Q1 [Mar May] 1[0 2.25] 1.0 [0 3] 2.0 [1 3] 2.0 [1 3.5] Q2 [June Aug] 1.5 [0 2] 1.0 [0 2] 2.0 [1 3] 2.0 [1 3] Q3 [Sept Nov] 1.0 [0 2] 1.0 [0 2] 2.0 [1 3] 2.0 [2 3] Q4 [Dec Feb] 1.0 [0 2] 1.0 [0 2] 1.5 [ ] 1.0 [0 3] χ df P value MRC Medical research council in the incidence of GBS. [10] In our study too, there was a trend toward higher number of patients being admitted from March to August (59.3%). The seasonal increase in the incidence of GBS has been attributed to the increased peaks of gastrointestinal and respiratory tract infections in these seasons. However, none of these studies evaluated the association of the season with the severity of illness or rapidity of recovery. About a third of patients of GBS require mechanical ventilation. [11] A poor disability grade and areflexia on admission predict respiratory failure. [12] In our study, in mechanically ventilated GBS patients, the best muscle power at admission in the limbs was 3/5. The need for invasive mechanical ventilation is predicted by several factors. They include rapidly progressive motor weakness, involvement of the peripheral limb and the axial muscles, ineffective cough, bulbar muscle weakness, and a rapid decrease in vital capacity. [13] Most of the published studies attempted to predict the need for mechanical ventilation, which is essential to plan the ICU usage, but none of them analyzed the predictive factors for the duration of mechanical ventilation and hospital stay. The latter variations are as important as the former, given their impact on the hospital resource utilization. Among the electrophysiological correlates, axonopathy has been an important factor that has been shown to be predictive of poor neurological outcome and prolonged respiratory paralysis. [14] In our study, the number of Table 4: ICU factors likely to affect the duration of ventilation and ICU stay Parameters Q1 [n=52] Q2 [n=40] Q3 [n=35] Q4 [n=28] χ 2 Significance Respiratory complications (df=3) 0.4 Cardiovascular complications (df=3) 0.2 Nutritional complications (df=3) 0.7 Metabolic complications (df=3) 0.9 Infectious complications (df=3) 0.9 ICU Intensive care unit Table 5: Recovery characteristics of patients admitted in different seasons Parameter Q1 [n=52] Q2 [n=40] Q3 [n=35] Q4 [n=28] P value Post hoc comparisons Duration of mechanical ventilation (days) 23±20 36±34 27±22 38± Q1 vs Q2=0.03 Q1 vs Q4=0.02 Q3 vs Q2=0.16 Q3 vs Q4=0.11 Duration of ICU stay (days) 27±22 40±37 31±23 43± Q1 vs Q2=0.03 Q1 vs Q4=0.02 Q3 vs Q2=0.16 Q3 vs Q4=0.11 Duration of hospital stay (days) Time to walk from onset of illness (days) [n=83; 28, 15, 21, 19 respectively] Time to recovery from treatment (days) 42±28 55±44 47±34 72± Q1 vs Q2=0.13 Q1 vs Q4=0.002 Q3 vs Q2=0.40 Q3 vs Q4= ±46 62±40 88± ± Q1 vs Q2=0. 8 Q1 vs Q4=0.02 Q3 vs Q2=0.45 Q3 vs Q4= ±14 29±34 18±14 29± Q1 vs Q2=0.008 Q1 vs Q4=0.02 Q3 vs Q2=0.05 Q3 vs Q4= Neurology India Jul-Aug 2013 Vol 61 Issue 4

5 r patients with diagnosis of primary axonopathy and also the number of patients with unexcitable neurons did not differ across the four seasons. This implies that additional factors like season also play a role in remyelination and recovery from GBS. In the present study, we examined several other factors that could potentially influence the recovery from GBS. But none of these variables turned out to be significantly different between seasons, which suggests that the observed difference is truly related to the effect of season on recovery from GBS. In particular, we examined the impact of the duration of specific treatment prior to commencement of ventilation on the outcome variables and found no difference across the groups [Table 2]. Therefore, we believe that occurrence of GBS in a particular season might itself independently predict the rate of recovery and, thereby, the duration of mechanical ventilation and ICU and hospital stay. In any demyelinating disease, the rate of clinical recovery is dependent on the rate of remyelination and proliferation of Schwann cells which produce cytokines and neural growth factors that facilitate regeneration of nerves. [15] At this time, we are unable to establish the mechanism of association between the season and outcome in GBS patients with certainty. One possible hypothesis is that it results from differential response of the immune system to varying offending agents in different seasons. This needs to be examined in prospective studies that focus on the association between the immunological response to the prodromal illnesses and the recovery from GBS. The second possibility is that the rate of remyelination of the damaged neurons could be influenced by some seasonal factor, which requires further investigation. In one experimental study, a shorter day length (as in winter) has been shown to result in slower myelination and the longer day lengths (as in summer) result in faster myelination through their influence on the rate of cellular proliferation. [16] Thirdly, the virulence of the causative viral/bacterial infection could be different in different seasons and, hence, though the antecedent illness is similar, the response of the host could be varied. Earlier studies have debated on the relative efficacies of various modalities of definitive therapies in GBS. The current consensus is that corticosteroids are ineffective, [17,18] while plasmapheresis and IVIG are equally effective. [19] All patients in our study received at least one form of definitive therapy and the distribution of therapies was uniform across all seasons. Therefore, the differences in the rate of recovery observed are unlikely to have been influenced by the modality of definitive therapy. Our results point out that acute inflammatory demyelinating polyneuropathy, which is considered a Neurology India Jul-Aug 2013 Vol 61 Issue 4 single homogeneous entity, is possibly a conglomeration of multiple immune pathological variants. In this study, we excluded patients who died in the ICU from analysis, as outcome parameters of the study cannot be assessed with certainty in these patients. We, however, performed an analysis by including the patients who died in the ICU and the results still remained the same. In the present study, we limited our analysis of outcome only to those patients who required mechanical ventilation. It would be interesting to extend the study of recovery characteristics to non ventilated patients too. If similar results are seen, it would further strengthen our observation and would call for a systematic analysis of the relationship between the seasonal etiological factors and the duration of recovery from GBS. Secondly, it is possible that the variations of seasons in Bangalore (India) might be different from those elsewhere. That makes it difficult to extrapolate the observations of this study to other parts of the world. Thirdly, while it may be true that management of patients with GBS may not change with this new knowledge, this knowledge will help to facilitate ICU resource management for the hospitals. Fourthly, evolution of ICU care over the last decade is unlikely to have impacted the findings in our study as the definitive therapy of GBS and ICU care was uniform across various seasons in any given year. The influence of ICU acquired weakness which can affect the recovery parameters could not be assessed due to the retrospective nature of this study; however, its impact on already demyelinated nerves is unknown. Lastly, this study suffers from the consequences associated with any retrospective study. A well designed prospective, multicenter study with different climatic conditions and standardized definitive care with either IVIG or plasmapheresis and similar ICU management protocol will establish these findings across various geographic locations. References 1. The prognosis and main prognostic indicators of Guillain Barré syndrome. A multicentre prospective study of 297 patients. The Italian Guillain Barré study group. Brain 1996;119: Hadden RD, Karch H, Hartung HP, Zielasek J, Weissbrich B, Schubert J, et al. Preceding infections, immune factors, and outcome in Guillain Barré syndrome. Neurology 2001;56: Netto AB, Taly AB, Kulkarni GB, Uma Maheshwara Rao GS, Rao S. Prognosis of patients with Guillain Barré syndrome requiring mechanical ventilation. Neurol India 2011;59: Van Doorn PA. Treatment of Guillain Barré syndrome and CIDP. J Peripher Nerv Syst 2005;10: Van Koningsveld R, Steyerberg EW, Hughes RA, Swan AV, van Doorn PA, Jacobs BC. A clinical prognostic scoring system for Guillain Barré syndrome. Lancet Neurol 2007;6: Netto AB, Taly AB, Kulkarni GB, Rao UG, Rao S. Mortality in mechanically ventilated patients of Guillain Barré Syndrome. Ann Indian Acad Neurol 2011;14: Zaheer M, Naeem M, Nasrullah M. Seasonal variation and sex distribution in patients with Guillain Barré syndrome. Pak J Neurol Sci 2008;3:

6 8. Hughes RA, Rees JH. Clinical and epidemiologic features of Guillain Barré syndrome. The J Infect Dis 1997;176:S Baoxun Z, Yinchang Y, Huifen H, Xiuqin L. Acute polyradiculitis (Guillain Barré syndrome): An epidemiological study of 156 cases observed in Beijing. Ann Neurol 1981;9: A prospective study on the incidence and prognosis of Guillain Barré syndrome in Emilia Romagna region, Italy ( ). Emilia Romagna study group on clinical and epidemiological problems in neurology. Neurology 1997;48: Chalela JA. Pearls and pitfalls in the intensive care management of Guillain Barré syndrome. Semin Neurol 2001;21: Cheng BC, Chang WN, Chang CS, Tsai NW, Chang CJ, Hung PL, et al. Predictive factors and long term outcome of respiratory failure after Guillain Barré syndrome. Am J Med Sci 2004;327: Orlikowski D, Prigent H, Sharshar T, Lofaso F, Raphael JC. Respiratory dysfunction in Guillain Barré Syndrome. Neurocrit Care 2004;1: Sundar U, Abraham E, Gharat A, Yeolekar ME, Trivedi T, Dwivedi N. Neuromuscular respiratory failure in Guillain Barré Syndrome: Evaluation of clinical and electrodiagnostic predictors. J Assoc Physicians India 2005;53: Lisak RP, Skundric D, Bealmear B, Ragheb S. The Role of Cytokines in Schwann Cell Damage, Protection, and Repair. J Infect Dis 1997;176(Suppl 2):S Spears N, Meyer JS, Whaling CS, Wade GN, Zucker I, Dark J. Long day lengths enhance myelination of midbrain and hindbrain regions of developing meadow voles. Brain Res Dev Brain Res 1990;55: Hughes RA, van Der Meche FG. Corticosteroids for treating Guillain Barré syndrome. Cochrane Database Syst Rev 2000;3:CD Hughes RA, Swan AV, van Doom PA. Corticosteroids for Guillain Barré syndrome. Cochrane Database Syst Rev 2010;(2):CD Randomised trial of plasma exchange, intravenous immunoglobulin, and combined treatments in Guillain Barré syndrome. Plasma Exchange/Sandoglobulin Guillain Barré Syndrome Trial Group. Lancet 1997;349: How to cite this article: Sriganesh K, Netto A, Kulkarni GB, Taly AB, Umamaheswara Rao GS. Seasonal variation in the clinical recovery of patients with Guillain Barré syndrome requiring mechanical ventilation. Neurol India 2013;61: Source of Support: Nil, Conflict of Interest: None declared. Author Help: Online submission of the manuscripts Articles can be submitted online from For online submission, the articles should be prepared in two files (first page file and article file). Images should be submitted separately. 1) First Page File: Prepare the title page, covering letter, acknowledgement etc. using a word processor program. All information related to your identity should be included here. Use text/rtf/doc/pdf files. Do not zip the files. 2) Article File: The main text of the article, beginning with the Abstract to References (including tables) should be in this file. Do not include any information (such as acknowledgement, your names in page headers etc.) in this file. Use text/rtf/doc/pdf files. Do not zip the files. Limit the file size to 1 MB. Do not incorporate images in the file. If file size is large, graphs can be submitted separately as images, without their being incorporated in the article file. This will reduce the size of the file. 3) Images: Submit good quality color images. Each image should be less than 4 MB in size. The size of the image can be reduced by decreasing the actual height and width of the images (keep up to about 6 inches and up to about 1800 x 1200 pixels). JPEG is the most suitable file format. The image quality should be good enough to judge the scientific value of the image. For the purpose of printing, always retain a good quality, high resolution image. This high resolution image should be sent to the editorial office at the time of sending a revised article. 4) Legends: Legends for the figures/images should be included at the end of the article file. 354 Neurology India Jul-Aug 2013 Vol 61 Issue 4